Made available in DSpace on 2014-06-11T19:22:32Z (GMT). No. of bitstreams: 0 Previous issue date: 2011-12-02Bitstream added on 2014-06-13T19:08:06Z : No. of bitstreams: 1 mantovani_wa_me_ilha.pdf: 572859 bytes, checksum: 6f6dbbff3d2181a18b8a8a253a4fc213 (MD5) / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Com o surgimento de novas tecnologias, a eletrônica de potência passou a ser utilizada cada vez mais, resultando no aumento das cargas não-lineares e na adoção de dispositivos que provocam distorções harmônicas na rede elétrica. Essas distorções harmônicas geram maiores custos às empresas distribuidoras de energia elétrica, como a troca de cabos e redimensionamento de transformadores, que podem sofrer superaquecimento ou sobrecarga. Estes fenômenos prejudicam o fornecimento de energia e podem gerar futuras taxações a estas empresas, uma vez que já existe a discussão de normas que limitam o nível de distorções harmônicas presentes nos sistemas de distribuição. Uma vez constatada uma violação dos limites estabelecidos, torna-se necessária a adoção de uma sistemática para mitigação dos efeitos resultantes, que na maioria dos casos está relacionada a instalação de filtros harmônicos, sejam de natureza passiva, ativa ou híbridos. Por outro lado, em decorrência desta medida, surge naturalmente a questão vinculada a responsabilidade financeira sobre os investimentos relacionados com os procedimentos de mitigação a serem implementados. Para a realização do compartilhamento de responsabilidades, é necessário determinar a parcela de distorção harmônica total relativa somente à não-linearidade da carga. Até a presente dissertação, a única forma de se obter tal valor era desacoplando a carga do sistema e alimentando-a com uma tensão senoidal, tornando-se um processo difícil e até inviável. A proposta deste trabalho é a utilização de redes neurais recorrentes na determinação da característica não-linear da carga através da medida da tensão no PAC e da corrente no ramo da carga em questão. A partir de um processo de treinamento, a rede neural simula a admitância da carga através... / With the advent of new technologies, the power electronics has become increasingly used, resulting in increased non-linear loads and the adoption of devices that cause harmonic distortion on the grid. These harmonic distortions generate higher costs for electricity distribution companies, such as replacing cables and transformer sizing, which can overheat or overload. These problems affect the power supply and can create future taxation of these companies. There is already discussion of rules limiting the level of harmonic distortion present in distribution systems. When is found a violation of limits, becomes necessary to adopt a systematic approach to mitigating the effects arising, which in most cases is related to installation of harmonic filters, are such as passive, active or hybrid. On the other hand, as a result of this action, the question naturally arises linked to financial liability on investments related to the mitigation procedures to be implemented. For the realization of shared responsibility, is necessary to determine the share of total harmonic distortion relative just to the non-linearity of the load. Until the present work, the only way to obtain that value was decoupling the system load and feeding it with a sinusoidal voltage, making it a difficult process and even infeasible. The purpose of this work is the use of recurrent neural networks in determining the non-linear characteristic of the load by measuring the voltage at the PCC and the current in branch of the load in question. From a training process, the neural network simulates the admittance of the load by adjusting their weights. Then these weights are transferred to a second neural network simulation, which when applied to enter a purely sinusoidal voltage, a current referring just to non-linear characteristic of the load is obtained

Orientador: Luiz Fernando Bovolato / Banca: Luís Carlos Origa de Oliveira / Banca: Olivio Carlos Nascimento Souto / Resumo: Com o surgimento de novas tecnologias, a eletrônica de potência passou a ser utilizada cada vez mais, resultando no aumento das cargas não-lineares e na adoção de dispositivos que provocam distorções harmônicas na rede elétrica. Essas distorções harmônicas geram maiores custos às empresas distribuidoras de energia elétrica, como a troca de cabos e redimensionamento de transformadores, que podem sofrer superaquecimento ou sobrecarga. Estes fenômenos prejudicam o fornecimento de energia e podem gerar futuras taxações a estas empresas, uma vez que já existe a discussão de normas que limitam o nível de distorções harmônicas presentes nos sistemas de distribuição. Uma vez constatada uma violação dos limites estabelecidos, torna-se necessária a adoção de uma sistemática para mitigação dos efeitos resultantes, que na maioria dos casos está relacionada a instalação de filtros harmônicos, sejam de natureza passiva, ativa ou híbridos. Por outro lado, em decorrência desta medida, surge naturalmente a questão vinculada a responsabilidade financeira sobre os investimentos relacionados com os procedimentos de mitigação a serem implementados. Para a realização do compartilhamento de responsabilidades, é necessário determinar a parcela de distorção harmônica total relativa somente à não-linearidade da carga. Até a presente dissertação, a única forma de se obter tal valor era desacoplando a carga do sistema e alimentando-a com uma tensão senoidal, tornando-se um processo difícil e até inviável. A proposta deste trabalho é a utilização de redes neurais recorrentes na determinação da característica não-linear da carga através da medida da tensão no PAC e da corrente no ramo da carga em questão. A partir de um processo de treinamento, a rede neural simula a admitância da carga através... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: With the advent of new technologies, the power electronics has become increasingly used, resulting in increased non-linear loads and the adoption of devices that cause harmonic distortion on the grid. These harmonic distortions generate higher costs for electricity distribution companies, such as replacing cables and transformer sizing, which can overheat or overload. These problems affect the power supply and can create future taxation of these companies. There is already discussion of rules limiting the level of harmonic distortion present in distribution systems. When is found a violation of limits, becomes necessary to adopt a systematic approach to mitigating the effects arising, which in most cases is related to installation of harmonic filters, are such as passive, active or hybrid. On the other hand, as a result of this action, the question naturally arises linked to financial liability on investments related to the mitigation procedures to be implemented. For the realization of shared responsibility, is necessary to determine the share of total harmonic distortion relative just to the non-linearity of the load. Until the present work, the only way to obtain that value was decoupling the system load and feeding it with a sinusoidal voltage, making it a difficult process and even infeasible. The purpose of this work is the use of recurrent neural networks in determining the non-linear characteristic of the load by measuring the voltage at the PCC and the current in branch of the load in question. From a training process, the neural network simulates the admittance of the load by adjusting their weights. Then these weights are transferred to a second neural network simulation, which when applied to enter a purely sinusoidal voltage, a current referring just to non-linear characteristic of the load is obtained / Mestre

Wave energy is a renewable energy source with a large potential to contribute to the world's electricity production. There exist several technologies on how to convert the energy in the ocean waves into electric energy. The wave energy converter (WEC) presented in this thesis is based on a linear synchronous generator. The generator is placed on the seabed and driven by a point absorbing buoy on the ocean surface. Instead of having one large unit, several smaller units are interconnected to increase the total installed power. To convert and interconnect the power from the generators, marine substations are used. The marine substations are placed on the seabed and convert the fluctuating AC from the generators into an AC suitable for grid connection. The work presented in the thesis focuses on the first steps in the electric energy conversion, converting the voltage out from the generators into DC, which have an impact on the WEC's ability to absorb and produce power. The purpose has been to investigate how the generator will operate when it is subjected to different load cases and to obtain guidelines on how future systems could be improved. Offshore experiments and simulations have been done on full scale generators connected to four different loads, i.e. one linear resistive load and three different non-linear loads representing different cases for grid connected WECs. The results show that the power can be controlled and optimized by choosing a suitable system for the WEC. It is not obvious which kind of system is the most preferable, since there are many different parameters that have an impact on the system performance, such as the size of the buoy, how the generator is designed, the number of WECs, the highest allowed complexity of the system, costs and so on. Therefore, the design of the electrical system should preferably be carried out in parallel with the design of the WEC in order to achieve an efficient system. / <p>Felaktigt tryckt som Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 727</p>

The technology of power electronics is used increasingly during recent years, and the electronic power facilities are used more and more in the power system. The non-linear electronic loads produce heavy harmonic currents and could significantly degrade the power quality. Nonlinear loads, including the un-interruptible power supply, motor control and converter, etc, are important equipment in a modern factory, however, these nonlinear loads could lead to power facility malfunction and capacitor damage. The harmonics would eventually cause severe unexpected capital loss. Identification of harmonic sources location becomes an important study for power quality. An effective tool is thus helpful for the harmonic source locating. This paper proposes a method to deal with the harmonic sources and location detection in the power system by using the artificial neural network (ANN). The non-linear loading characteristics are studied by the power flow analysis, and then the proposed methodology uses the Probabilistic Neural Networks¡]PNN¡^and wavelet-probabilistic network (WPN) for harmonic source locating. An IEEE 14-bus power system is used for study to show the effectiveness of the proposed approach.

Στην παρούσα διπλωματική εργασία εξετάζονται δύο υβριδικές διατάξεις με παρόμοια συνδεσμολογία, οι οποίες χρησιμοποιούνται για το φιλτράρισμα αρμονικών σε περιπτώσεις μη γραμμικών φορτίων. Συγκεκριμένα, περιγράφονται τα χαρακτηριστικά τους, τα κυκλώματα ελέγχου τους και αναλύονται λεπτομερώς τα αποτελέσματα τους στο φιλτράρισμα. Ιδιαίτερη βαρύτητα δίνεται στην έννοια των αρμονικών και γίνεται αναφορά στις διαταραχές που προκαλούν τα μη γραμμικά φορτία σ’ένα σύστημα ισχύος. Επιπλέον, παρουσιάζονται όλες οι κατηγορίες των υπαρχόντων αρμονικών φίλτρων και γίνονται οι απαραίτητες συγκρίσεις ανάμεσα στα δύο υβριδικά φίλτρα και στους συμβατικούς αντισταθμιστές. Τέλος, αναφέρονται τα διεθνή πρότυπα των αρμονικών και γίνεται λόγος για τις τιμές των φίλτρων στην αγορά. / In this thesis, two hybrid devices for harmonic filtering due to non linear loads are being examined.Particularly, their characteristics and their control systems are being described and their filtering performance is analysed in detail. Attention is paid to the term of Harmonics and the distortion that the non linear loads cause to power systems. Moreover, all kinds of harmonic filters are presented and compared. Finally, there is a short reference to the international standards of harmonics and the prices of those devices in the market.

Αυτή η διπλωματική εργασία εξετάζει θέματα που αφορούν την παρεχόμενη ποιότητα ισχύος στα σύγχρονα ηλεκτρικά συστήματα. Θα εξετάσουμε την απόδοση ενός στατικού σύγχρονου αντισταθμιστή STATCOM, που λειτουργεί ως ρυθμιστής τάσης, σε ένα σύστημα αυτοδιεγειρόμενης επαγωγικής γεννήτριας που τροφοδοτεί μη γραμμικά φορτία. Εκτός της θεωρητικής ανάλυσης των συνιστωσών του εξεταζόμενου συστήματος, θα πραγματοποιήσουμε προσομοίωσή του, με τη βοήθεια του λογισμικού PSCAD/EMTDC, με σκοπό την εξαγωγή χρήσιμων συμπερασμάτων από τα παραγόμενα αποτελέσματα. Ειδικό ενδιαφέρον θα δοθεί εκτός από την απόδοση στη ρύθμιση της τάσης και στη συνολική αρμονική παραμόρφωση της τάσης και την επίδραση των αρμονικών στην ποιότητα ισχύος. / This diploma thesis studies issues concerning the provided power quality in modern power systems. We will examine the performance of a STATCOM (Static synchronous Compensator) based voltage regulator in a, self excited induction generator (SEIG) feeding non-linear loads system. Besides the theoretical analysis of each component of the examined system, we will simulate it with the help of a software program called PSCAD/EMTDC in order to get useful conclusions through the provided results. Specific attention will be given, besides the efficiency in voltage regulation, in total harmonic distortion and the effect harmonics cause in power quality.

Made available in DSpace on 2014-07-29T15:08:23Z (GMT). No. of bitstreams: 1 Dissertacao Daywes neto.pdf: 2344680 bytes, checksum: b21938ab4c7959bef7e6333564003d70 (MD5) Previous issue date: 2010-03-05 / This work analyzes the performance of single-phase and three-phase power meters, of the electromechanical and electronic types, when they are working with linear and nonlinear loads present in residence, commerce and industry. It differs from related studies mainly for making use of actual loads encountered in day-to-day, rather than sources of electronic charges as has been observed in another papers. The comparison of results is made taking as standard a meter power of high precision developed through virtual instrumentation / O presente trabalho analisa o comportamento de medidores monofásicos e trifásicos de energia elétrica dos tipos eletromecânico e eletrônico, quando os mesmos são submetidos a ambientes de cargas lineares e não-lineares, dos tipos residencial, comercial e industrial. Diferencia-se dos estudos correlatos principalmente por fazer uso de cargas reais encontradas no dia-a-dia, em vez de fontes de cargas eletrônicas como se tem observado no estado da arte. A comparação de resultados é feita tomando como base um padrão de energia de alta precisão desenvolvido por meio de instrumentação virtual

Submitted by Renata Lopes (renatasil82@gmail.com) on 2016-06-09T12:11:11Z No. of bitstreams: 1 pedrodacruzloureiro.pdf: 1767688 bytes, checksum: 1fa1e4fbfaa6feaf5a5c88ea70df09d6 (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2016-07-13T13:31:22Z (GMT) No. of bitstreams: 1 pedrodacruzloureiro.pdf: 1767688 bytes, checksum: 1fa1e4fbfaa6feaf5a5c88ea70df09d6 (MD5) / Made available in DSpace on 2016-07-13T13:31:22Z (GMT). No. of bitstreams: 1 pedrodacruzloureiro.pdf: 1767688 bytes, checksum: 1fa1e4fbfaa6feaf5a5c88ea70df09d6 (MD5) Previous issue date: 2012-04-16 / CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Neste trabalho é proposta a aplicação de redes neurais artificiais para ajuste de parâmetros de um compensador estático de reativos, para controle de tensão e harmônicos. Devido à intensa produção de correntes harmônicas e possíveis afundamentos de tensão em instalações industriais como o forno a arco, é necessário um sistema de controle eficiente e robusto. Além disso, os sistemas elétricos de potência se encontram em um cenário com a presença cada vez maior de geração distribuída, cargas não-lineares e forte tendência à operação no contexto das smartgrids e microgrids. Sendo assim, o suporte de reativos deve ser adequado a esses sistemas, podendo atuar de forma rápida, precisa e confiável. Uma possível solução é a utilização de um compensador estático de reativos (CER) com função adicional de filtragem no ponto onde se deseja controlar a tensão e a distorção harmônica. Entretanto, para o correto funcionamento, é necessário um sistema preciso para o ajuste dos parâmetros do CER, ou seja, determinar os ângulos de disparo dos tiristores e o número de bancos de capacitores a serem ligados. Neste trabalho é proposta uma estratégia de controle via redes neurais artificiais, treinadas para o reconhecimento de padrões de operação em regime permanente e definição da configuração do CER, conferindo inteligência ao equipamento. Os desenvolvimentos propostos foram implementados no ambiente MatLab®. A validação do método é feita através de simulações em sistemas-teste, presentes na literatura técnica, utilizando o fluxo de potência pelo método de injeção de correntes trifásico harmônico. Os resultados obtidos mostram as vantagens da utilização da estratégia proposta. / In this work, an artificial neural network-based static var compensator tuning is proposed for voltage and harmonic distortion control. Due to intense harmonic current injection and possible voltage sags produced by industrial facilities such as arc furnaces, an efficient robust control system is needed. Besides, electrical power systems face a new scenario with high penetration of distributed generation and non-linear loads and increased smart grid and microgrid trends. Therefore, the available reactive power sources must be able to provide system control in order to operate the system in a fast, accurate and reliable way. The application of a static var compensator (SVC) with additional filtering function at the controlled node is a possible solution. However, a precise SVC parameters tuning is needed, in order to make the system to work properly. In this work, a control strategy based on artificial neural networks is proposed. The neural networks are trained to recognize steadystate operating patterns and give the SVC adjustment. The proposed technique was implemented in the MatLab® environment. The methodology is validated by simulations in test-systems available in technical literature, using the three-phase harmonic current injection method power flow. Results show the advantages of the proposed methodology.