Made available in DSpace on 2014-06-11T19:22:32Z (GMT). No. of bitstreams: 0 Previous issue date: 2012-03-01Bitstream added on 2014-06-13T18:08:27Z : No. of bitstreams: 1 faria_lt_me_ilha.pdf: 1374393 bytes, checksum: 021c232f9110d70353e324741a8e1388 (MD5) / Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) / As perdas de energia elétrica por fraudes, ligações clandestinas ou erro na medição são denominadas Perdas Não-Técnicas ou Perdas Comerciais e seu combate tem sido prioridade quer por empresas concessionárias de distribuição de energia elétrica, quer por órgãos reguladores. Nesse contexto, neste trabalho, implementa-se computacionalmente um sistema inteligente híbrido intercomunicativo específico que se baseia no emprego de diferentes técnicas oriundas da área de sistemas inteligentes tais como redes neurais e lógica fuzzy em módulos independentes e que se comunicam entre si. O sistema é baseado em três pilares: extração automática de conhecimento a partir da base de dados da concessionária, incorporação na metodologia do conhecimento e experiência de especialistas e, em último, consultas na base de dados por características específicas de cada cliente. A metodologia utiliza simultaneamete inúmeros dados reais de entrada de natureza diversa e combina várias técnicas a fim de verificar o risco percentual de cada cliente de possuir alguma anomalia que implique em perda comercial. Além dos dados cadastrais e do histórico de consumo mensal dos clientes, comumente utilizados pelos trabalhos orientados à detecção de perdas comerciais, a metodologia proposta utilizou também dados adicionais tais como a lista de nomes e de atividades suspeitas. A utilização de dados adicionais possibilitou uma melhoria na detecção de clientes anômalos, grande parte dos quais seriam possivelmente considerados normais pelos trabalhos da literatura avaliada. Neste trabalho, pretende-se detectar as perdas comerciais de maneira mais rápida e precisa possível. Investigações adicionais devem ser feitas posteriormente para encontrar quais são as causas que culminaram em altas perdas comerciais em... / Electrical energy losses due to theft, fraud or error in the measurement are called Non-Technical Losses and their reduction has been a priority utilities power and by regulators. In this context, this paper presents the computational implementation of an intelligent hybrid system that combines techniques such as neural networks and fuzzy logic. The system is based on three pillars: knowledge extraction from the database utility, incorporating the methodology of knowledge and experience of experts and queries the database for specific features of each client. The methodology uses several simultaneous input of diverse nature and combines several techniques to verify the percentage risk of each customer to have some problem to configure non-technical losses as fraud, defective in the measurement system. In this paper, the main objective is to locate the focus of the problem more quickly. Further investigations should be made later to find what are the causes that resulted in high nontechnical losses in a feeder or in a specific region. This work covers the registered customers at utilities power, especially residential... (Complete abstract click electronic access below)

Made available in DSpace on 2014-06-11T19:22:33Z (GMT). No. of bitstreams: 0 Previous issue date: 2013-03-20Bitstream added on 2014-06-13T20:29:04Z : No. of bitstreams: 1 lima_fpa_me_ilha.pdf: 1444962 bytes, checksum: f814ed07710ed9ad108eae8a90aec20d (MD5) / Neste trabalho apresenta-se um método para detecção e classificação de distúrbios de tensão em sistemas de distribuição de energia elétrica. A partir das oscilografias de tensão, medidas na subestação, aplica-se o algoritmo de seleção negativa de um sistema imunológico artificial para diferenciar os sinais entre próprios e não-próprios. Os sinais classificados como próprios representam a operação normal do sistema, isto é, alimentador sem a presença de distúrbio, e os classificados como não-próprios são sinais onde há a presença de uma anormalidade. A análise dos sinais é realizada através do janelamento das oscilografias onde são realizadas comparações entre os sinais e os detectores previamente criados, avaliando a afinidade entre as janelas. Caso a afinidade entre os sinais ultrapasse um limiar preestabelecido pelo operador, é encontrado um casamento, e o sinal é classificado. Com esta proposta busca-se a concepção de um método que possa ser modificado facilmente, para atender a permanente evolução das tecnologias das subestações. Para validar e avaliar o desempenho do método foram realizados testes com quatro sistemas de distribuição de energia elétrica, sendo três sistemas testes de 5, 33 e 84 barras e um sistema real de 134 barras / This work presents a method to detect and classify voltage disturbances in electrical distribution systems. The Negative Selection Algorithm of an artificial immunological system is applied to the voltage oscillographs, measured on the substation, to distinguish the signals as self or nonself. The signals classified as self represent the normal operation of the system, i.e. the feeder without disturbance, and those classified as nonself are the signals where there are some abnormalities. The signal analysis is performed by windowing the oscillographs comparing the signals and the detectors previously created, evaluating the affinity with the windows. If the affinity with the signals overpasses a predefined limit a matching is found and the signal is classified. This proposal is a method that can be easily modified to attend the fast evolution of the substation technologies. To validate the performance tests were done to four different distribution systems, including one real system

Made available in DSpace on 2014-06-11T19:22:35Z (GMT). No. of bitstreams: 0 Previous issue date: 2006-04-26Bitstream added on 2014-06-13T19:48:59Z : No. of bitstreams: 1 pantuzi_av_me_ilha.pdf: 1781418 bytes, checksum: 8c96b8f3bf6cc971c43ea4aa045c10fb (MD5) / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Neste trabalho investigou-se o desempenho de um algoritmo de cálculo de fluxo de potência trifásico baseado em um método de varredura chamado backward forward sweep. O estudo desenvolveu-se para sistemas de distribuição de energia elétrica radiais e/ou radiais fracamente malhados, sendo que o foco principal foi a avaliação do processo de convergência e exatidão do método de fluxo de potência. Para tal finalidade investigou-se variadas situações de redes, tais como: barras conectadas com linhas longas e linhas curtas, presença de reguladores de tensão, presença de geração distribuída, presença de linhas com representação shunt, sistemas com carregamento leve, carregamento médio e carregamento pesado, desequilíbrio de cargas entre as fases (a, b e c), influência de modelos de cargas (potência constante, admitância constante e corrente constante) etc. O método backward-forward sweep é bem conhecido e discutido na literatura, porém uma análise de desempenho com todos os detalhes de modelos existentes nas redes de distribuição não é encontrada. Encontram-se análises com modelos isolados, ou seja, considerando somente redes fracamente malhadas, ou com barras com tensão controlada, ou com modelos de transformadores etc. Muitos destes estudos apresentaram resultados apenas para redes de pequeno porte, pois na maioria dos casos o objetivo principal era validar os modelos. Neste trabalho buscou-se desenvolver um estudo completo de desempenho e confiabilidade de uma implementação do método backward-forward sweep. Assim foram utilizadas redes de distribuição reais e fictícias encontradas na literatura, com as mais variadas topologias, incluindo ramais monofásicos (MRT), diferentes níveis de carregamento, cargas desequilibradas e diversos modelos de componentes de rede. / At this work an algorithm of calculation on three-phase power flow was investigated based on a sweeping method called backward forward sweep. The study was developed for radial electric power distribution networks and/or weakly interconnected radial ones, and the main focus was the evaluation of the convergence process and accuracy of the method of power flow. For such a purpose it was investigated varied situations of networks, such as: connected bars with long lines and short lines, voltage regulators presence, distributed generation presence, lines with representation shunt, systems with light, medium and heavy loading, unbalanced of loads among the phases (a, b and c) and influence of loads models (constant power, constant current and constant admittance). The method backward-forward sweep is very known and discussed in the literature, however an analysis with all the details of existent models in the distribution networks is not found. They are analyses with isolated models, in other words, considering only networks weakly interconnected, or with bars with controlled voltage, or with models of transformers. Many of these studies presented resulted just for networks of small load, because in most of the cases the main objective was to validate the models. In this work it was looked for to develop a complete study and reliability of an implementation of the method backward-forward sweep. Like this real and fictitious distribution networks were used found in the literature, with the most varied topologies, including monophase extensions (MRT), different loading levels, unbalanced loads and several models of networks components.

Orientador: José Roberto Sanches Mantovani / Coorientador: Marcos Júlio Rider Flores / Banca: Rubén Augusto Romero Lázaro / Banca: Carlos Roberto Minussi / Banca: João Bosco Augusto London Junior / Banca: Eduardo Nobuhiro Asada / Resumo: Neste trabalho é apresentado um Estimador de Estado para Sistemas de Distribuição de Energia Elétrica desbalanceados que usa uma nova formulação do método de fluxo de carga de varredura Distflow junto com o método dos Mínimos Quadrados Ponderados para determinar o ponto de operação mais provável do sistema. Através desta formulação, é possível modelar facilmente as medidas de magnitude de corrente, assim como as medidas convencionais de magnitude de tensão, as injeções e os fluxos de potências ativa e reativa. Adicionalmente, é definido um novo conjunto de variáveis de estado que permite reduzir o número de incógnitas do problema e, portanto, reduzir o número de medidas necessárias para realizar a estimação de estado. Assim, o estimador de estado proposto busca demonstrar que, mediante o uso da nova formulação, é possível determinar satisfatoriamente o ponto de operação dos Sistema de Distribuição de Energia Elétrica com altos níveis de automação, usando um plano de medições com um número baixo de pseudomedidas. A metodologia proposta é testada e analisada através dos sistemas IEEE-37 e IEEE-123 disponíveis na literatura. / Abstract: This work presents a State Estimation method for Unbalanced Power Distribution Systems, using a novel Distflow equations and the Weighted Least Squares method, in order to obtain the most likely state of the network. Through this formulation, it is possible to consider the line current magnitude measurements, as well as, conventional measurements: bus voltage magnitudes, bus power injections and line power flows. In addition, a new set of state variables is defined, which allows to reduce the number of variables, therefore, to reduce the number of measurements needed for state estimation analysis. The proposed state estimator deal to demonstrate that using the novel formulation is possible to successfully determine the operating conditions of a Power Distribution System with high level of automation, using a set of measurements with a low number of pseudomeasurements. The proposed methodology is tested on the IEEE-37 bus system and the IEEE-123 bus system available in the literature. / Doutor

Orientador: Rubén Augusto Romero Lázaro / Banca: Sergio Azevedo de Oliveira / Banca: Carlos Roberto Mendonça da Rocha / Resumo: Neste trabalho propõe-se uma heurística especializada para a alocação ótima de bancos de capacitores para os sistemas de distribuição radiais, cujo objetivo é minimizar os custos devido às perdas de energia, sujeitas a certas restrições operativas do sistema elétrico de distribuição. Para realizar este objetivo foram desenvolvidos os seguintes tópicos: Um algoritmo heurístico construtivo, que tem como atividade principal a alocação de bancos de capacitores nas barras com demanda de potências reativas maiores e também procura uma solução inicial de minimização de custos das perdas de energia e uma metaheurística de busca em vizinhança variável descendente. Nessa metaheurística são implementadas estruturas de vizinhança adequadas para o problema de alocação de bancos de capacitores nas barras do sistema elétrico. A heurística especializada tem um novo conceito para a alocação ótima de bancos de capacitores ao desenvolver uma ferramenta amigável e poderosa de otimização no desenvolvimento de modelos matemáticos para os problemas de programação não linear inteiro misto. A heurística especializada foi implementada em linguagem de programação Fortran 90 e foram realizadas simulações com sistemas de distribuição radial de 33, 70 e 202 barras. Os resultados foram comparados com pesquisas que realizaram testes com esses sistemas elétricos. Assim, foi verificado um bom desempenho do algoritmo, pois gera soluções de boa qualidade atendendo a todas as restrições físicas e operacionais do problema / Abstract: In this work we proposes a specialized heuristic algorithm for optimal allocation of capacitor banks in radial distribution systems, whose goal is to minimize costs due to energy losses, subject to some operational restrictions of the electric distribution system. To accomplish this goal the following subjects were developed: A constructive heuristic algorithm, which has as primary task the allocation of capacitor banks in buses with the most high-demand for reactive power and also search for an initial solution to minimize energy losses costs and a meta-heuristic which uses variable neighborhood descend search, in which we implemented appropriate neighborhood structures for the allocation problem of capacitor banks in electric system buses. The specialized heuristic uses a new concept for the optimal allocation of capacitor banks to develop a powerful and user-friendly optimization tool for the development of mathematical models for mixed integer nonlinear programming problems. The specialized heuristic was implemented in the programming language Fortran 90 and simulations were performed with radial distribution systems of 33, 70 and 202 buses. The results were compared with other research works that also tested with these electrical systems. Thus, it was verified that the algorithm has a good performance because it generates good quality solutions meeting all the physical and operational constraints of the problem / Mestre

Orientador: Rubén Augusto Romero Lázaro / Banca: Fábio Bertequini Leão / Banca: Eduardo Nobuhiro Asada / Resumo: Neste trabalho é apresentada uma metodologia, baseada na meta-heurística de Busca Dispersa, para resolver o problema de alocação de bancos de capacitores em redes de distri- buição radiais. Foram feitas novas propostas, objetivando a melhoria do tempo de processa- mento e convergência do método. A função objetivo a ser minimizada incorpora os custos totais de investimentos (aquisição, instalação e manutenção) e de operação. Para o cálculo do ponto de operação do sistema foi implementado um algoritmo de Fluxo de Carga de varredu- ra. O método considera a instalação de bancos de capacitores fixos e chaveados. Os resultados obtidos com o algoritmo implementado reduziram as perdas elétricas e melhoraram o perfil de tensão do sistema, sendo comparados com os obtidos por outros dois métodos alternativos para o mesmo problema. A metodologia proposta foi implementada em linguagem MATLAB e testada em quatro sistemas presentes na literatura: o de 34, 136, 202 e 400 barras. A partir dos resultados foi verificado que a metodologia apresentou um bom desempenho / Abstract: This work proposes a methodology, based on the Scatter Search metaheuristic, to solve the problem of capacitor placement in radial distribution networks. New proposals were made to achieve a better convergence and computer performance of the method. The objec- tive function to be minimized incorporates the total value of investments (acquisition, installa- tion and maintenance of capacitor banks) and also operational costs. To obtain the operation state of the system, a specialized radial load flow algorithm is used. The method also consid- ers the installation of fixed and switched capacitor banks. The results show that loss reduction and voltage profile improvement can be achieved with the installation of capacitor banks in the network, they were compared with those obtained by two alternative methods for the same problem. The methodology was implemented in MATLAB and tested with four systems pre- sented in literature: the 34 bus, 136 bus, 202 bus and 400 bus systems. From the obtained re- sults it was verified that the methodology presented a good performance / Mestre

Thesis (MTech (Electrical Engineering))--Cape Peninsula University of Technology, 2009 / The industrial and population growth of a nation can cause power delivery problems to localized areas of a national grid through their increased demand for electrical energy. One reason for these power shortages is the insufficient current carrying capacity of existing high voltage alternating current, (HVAC), transmission lines supplying the area. High voltage direct current (HVDC) transmission lines are a possible solution as they provide better power delivery than HVAC lines. New or upgraded HVAC lines, or HVDC lines or combinations of HVAC and HVDC lines are possible solutions to improve power delivery. This research investigates the various line possibilities using theory. and cutting edge frequency and time domain software tools. The challenge is how to approach this problem. What methodology or structure should be used? Thus one of the contributions of this work is the development of a strategy (flow chart), for solving power delivery problems to localized areas of a national grid through individual or combinations (e.g. parallel operation) of HVAC and/or HVDC transmission lines. The main contribution is the evaluation of a HVDC system as a solution to overcoming power delivery shortages to a localized area of a national grid. Three different software packages (two industrial and one academic) namely, PSCAD/EMTDC (time domain), DlgSILENT PowerFactory (frequency domain) and MathCAD software are evaluated for their capability to perform the simulation studies necessary to prove the possible solutions given in the developed flow chart. The PSCAD/EMTDC software package is evaluated for integrated HVAC/HVDC load flow analyses, DlgSILENT for individual and parallel combinations of HVAC lines and MathCAD to prove hand calculations to software results. Five case studies are conducted. The first case study demonstrates a healthy system with no delivery shortcomings, the second case study portrays the delivery shortcoming due to increased localized area demand, and the remaining three case studies explore possible solutions to solve the problem. The first possible solution is to construct an identical HVAC line in parallel to the existing line.

Thesis (MTech (Electrical Engineering))--Cape Peninsula University of Technology, 2009 / Recently, due to concerns about the liberalization of electricity supply, deregulation and global impact on the environment, securing a reliable power supply has become an important social need worldwide To ensure this need is fulfilled, detailed investigations and developments are In progress on power distribution systems protection and the monitoring of apparatus which Is part of the thesis. The main objective of a protection schemes is mainly to keep the power system stable by isolating only the affected components or the section of the electricity network in which the fault has developed while allowing the rest of the network to continue operating. It is important to note that the protection equipment does not prevent faults from occurring, but it limits the damaging effect of the fault and protects other healthy equipment. This is only achieved if the protection system of the electrical network involved complies with the requirements and purpose of the electrical protection standards. These requirements include the Operational speed, Reliability, Security and Sensitivity. In conventional substations that are still existing Within the utilities networks, a number of long cables are then used to complete the links between substation equipment in order for them to communicate (hardwired). This method is uneconomical and is being phased out completely in the near future. Over the last few years a new standard for substation automation communication has been developed Within the International Electrotechnical Commission (IEC), the IEC 61850. This standard defines the integration requirements of multi-vendor compliant relays and other lED's for multiple protection schemes as well as control and automation techniques. In this particular thesis, Distribution protection is the area of interest, particularly the application of Time and Overcurrent protection schemes. A look into different protection alternatives and the application of new technologies for Electrical Power Distribution Systems that unify protection and control units so that they can be incorporated into Intelligent Substation as opposed to the most existing (conventional substation) is analyzed. The proposed algorithm has been verified through simulations of the CPUT and Eureka three phase power distribution systems. A testing Lab is also part of this thesis and Is meant for experiments as well as simulation performance in order to gain knowledge and skills for designing and engineer substations with lEG 61850 standards equiprnents. The results indicate that the reduction of copper wiring cable has increased and the communication speed has improved and simplified.

Thesis submitted in fulfilment of the requirements for the degree Master of Technology: Electrical Engineering in the Faculty of Engineering at the Cape Peninsula University of Technology 2014 / The research in this thesis deals with the application of Model Based Systems Engineering (MBSE) practices in the modelling of power systems. More particularly, we have presented the modelling hybrid photovoltaic wind integrated in an electrical distribution network using SysML (System Modelling Language) which is a modelling language in support of MBSE. MBSE refers to a formalised practice of systems development through the application of modelling principles, methods, languages and tools to the entire lifecycle of a system. Generally speaking, the modelling of power systems is performed using software such as Matlab Simulink, DigSilent, PowerWorld etc. These software programs allow modelling of a system considering only a specific viewpoint, depending on the objective that is to be assessed. The advantage of the SysML over the above mentioned modelling languages lies from the fact that SysML includes different viewpoints of a system. These views are known as the Four Pillars of SysML. Pillar One refers to the requirements of a system and includes all the functional and non-functional requirements. Pillar Two deals with the structure representation of a system by considering all its subsystems and their different connections. Pillar Three considers the behaviour of a system and includes its activities, sequences and different states. The last Pillar includes the detailed characteristics, physical laws and constraints on the system. The main objectives of this research are the development of models which will include: the system’s requirements; the system’s structure representation in term of different entities involved and the relationship between them; the system’s behaviours in terms of activities in different cases considered and transitions from one state to another as well as the interaction between the system and all the stakeholders. Keywords: Model Based Systems Engineering (MBSE), System Modelling Language (SysML), Renewable Energy systems, Hybrid power systems, photovoltaic systems, wind power systems.

Thesis submitted in fulfilment of the requirements for the degree Doctor of Technology: Electrical Engineering in the Faculty of Engineering at the Cape Peninsula University of Technology 2014 / Distributed generation (DG) has been reincarnated after its demise by centralised generation. While economy of scale and efficiency are the advantages of the latter, deregulation of the electricity market, environmental concerns and the need to arrest dwindling reserve margins have necessitated the rebirth of the former. Indeed, a full circle has therefore evolved with generation being ‘embedded’ in distribution systems and ‘dispersed’ around the system rather than being located and dispatched centrally or globally. This development is in tandem with the history of industrial revolutions that started from energy and moved through services and communication and back to energy. South Africa is not immune to the global energy, especially tertiary energy, challenge phenomenon. At the peak of the 2007-2008 energy crisis, her generation net reserve margin fell below 10% – well below conventional industry benchmark of at least 15%. Also South Africa is Africa’s largest emitter of CO2 contributing over 40% of Africa’s total CO2 emissions. Therefore, DG’s relevance to South Africa is quite obvious. However, DG integration into distribution networks leads to a number of challenges. For instance, with significant penetration of DG power flow reversal may be experienced and the distribution network will no longer be a passive circuit. This underscores the crucial role of ICT in active distribution network occasioned by DG and especially the emergent of “prosumerism” (a hitherto consumer also becoming a producer). Therefore, a smart grid and similar phrases have all been used to describe a “digitised” and intelligent version of the present-day power grid. There are immense benefits derivable from modelling and simulation. Consequently, a typical radial distribution network model has been developed to evaluate the considerable impacts of DG integration. The modelling and simulation of the network are accomplished using the DIgSILENT PowerFactory simulation package. Impacts of DG on voltage profile, fault level, voltage stability and protection coordination have been investigated and their possible mitigation measures proferred. The results reveal that for a particular DG type its impacts depend mainly on its capacity and point of connection relative to a given load type. Smart grid technology addresses some of these impacts through its inherent capability which includes peer-to-peer relay communication for protective devices on the distribution feeder as well as communication to the DG facility.