O aumento da demanda energética mundial e a necessidade de obtenção de energia por fontes renováveis abriram espaço para discussão sobre vantagens e desvantagens relacionadas às diversas tecnologias de geração existentes. Dados mostram um crescimento significativo do percentual da geração eólica no sistema energético mundial tornando factível o investimento em tecnologias que possam atender às diferenças de aplicação para cada região. Este trabalho apresenta inicialmente as principais características das arquiteturas construtivas dos geradores eólicos listando e apresentando seus principais equipamentos e funções, em seguida são apresentados métodos de controle e proteção aerodinâmica e as principais topologias aplicadas na geração eólica de energia. A discussão das vantagens e desvantagens de cada topologia apontou o gerador Doubly Fed Induction Generator (DFIG) aplicado em sistemas de velocidade variável como uma tecnologia que se destaca pelo seu alto nível de controle e pelo ganho relacionado ao conversor back-to-back de apenas 30% da capacidade de geração da máquina. No trabalho o controle do conversor back-to-back foi testado e validado por simulações e os ajustes dos controladores obtidos pelo método do modelo simétrico ótimo. Em seguida, são apresentados os dispositivos básicos de proteção correlacionados ao monitoramento de corrente, tensão e fluxo de potência e discutidas suas limitações na topologia com o gerador DFIG. No final do trabalho a proteção diferencial é apresentada e aplicada sobre o gerador DFIG e simulações de falta são realizadas em diferentes condições de funcionamento da máquina de modo a obter dados que possam apontar a melhor parametrização deste dispositivo de proteção. / The world energetic demand increase and the generation renewable necessity allowed the dialogue about the advantages and disadvantages of generations technologies existing. Many data shows an significant increase of wind generation inside the world energetic system allowing the investment in technologies to work at the characteristics existing in each place. This study shows the main constructive characteristics in the wind generation showing their main equipments and functions. After, the control and protection aerodynamic methods are shown together with the main topologies in the wind generation. The discussion about the advantages and disadvantages of this topologies showed the DFIG generator working in variable speeds systems like an efficient technology because your accurate control and the low cost of back-to-back frequency inverter. In this study the control of back-to-back converter was tested end validated and the drivers parameters was calculated with the method of the optimal symmetric model. then, are showed some basic protections devices linked with current, voltage and power flow monitoring and yours limitations are discussed in the DFIG topology. In the end the differential protection is showed and applied in the DFIG machine then fault simulations are make to obtain data to produce the best configuration parameters of differential protection.

A inserção de geradores distribuídos nas redes de distribuição de energia elétrica está aumentando consideravelmente devido à possibilidade de maior proximidade destes geradores com os mercados consumidores, fatores ambientais, aprimoramentos das tecnologias atuais de geração de energia elétrica e desenvolvimentos de outras novas, alterações de legislações e incentivos governamentais. Este novo cenário traz também a necessidade de maior atenção devido à possibilidade de ocorrência de ilhamentos não intencionais. Nestas situações, o gerador distribuído deve ser desconectado imediatamente após a detecção do evento. Métodos tradicionais, como relés de tensão e de frequência, mostram bom desempenho quando há um elevado desbalanço entre as potências geradas e as potências das cargas ilhadas. Nos casos de pequenos desbalanços de potência, o ilhamento pode não ser detectado por esses métodos. Neste contexto, este trabalho visa desenvolver um sistema de detecção de ilhamentos em sistemas de distribuição por meio da utilização de uma rede neural artificial do tipo Perceptron multicamadas. Utilizando-se uma rede neural devidamente treinada para o sistema elétrico estudado, obteve-se um bom desempenho na detecção de ilhamento, obtendo-se elevadas taxas de acerto. Além disso, o método proposto foi capaz de diferenciar corretamente situações normais de operação do sistema, tais como chaveamento de blocos de cargas, das situações de ilhamento. Os mesmos cenários foram testados utilizando as proteções convencionais de tensão, frequência e de taxa de variação de frequência, a fim de obter a comparação com o método proposto. Como resultado, o método baseado na rede neural apresentou desempenho superior aos demais sistemas de proteção, detectando o ilhamento dentro da faixa de 2 segundos em todos os casos simulados neste trabalho. / The penetration of distributed generators into the power distribution systems has been increasing significantly due to the possibility of generating power next to the consumers, environmental issues, improvement and development of the generation technologies, regulatory changes and governmental incentives. This new scenario requires special attention to the possibility of unintentional islanding. In this situation, the distributed generator must be disconnected immediately after an islanding occurrence. Traditional anti-islanding protection schemes, such as voltage and frequency-based relays, are effective to detect islanding if there is a big power imbalance between generation and load inside the island. On the other hand, they may fail if the power imbalance is small. In this context, this work presents a new islanding detection method based on a Perceptron Multi-layer neural network. By using a neural network properly trained for the distribution system under study, the method has presented good performance with high islanding detection successful rates. Moreover, the proposed method was capable of distinguish the islanding occurrence from normal transient operating conditions, such as load switching. The same operating scenarios were used to test voltage, frequency and rate of change of frequency protection in order to perform a comparison with the proposed method. As result, the method presented superior performance detecting the islanding faster than 2 seconds in all the cases simulated in this work.

The focus of this thesis is research on the performance of very low speed direct drive permanent magnet generators for energy conversion from marine and tidal currents. Various aspects involved in the design of these generators and their electromagnetic modelling using the finite element simulations are presented. For a detailed study, a 5 kW prototype generator has been designed and constructed based on finite element based simulations. Several experiments were conducted on the prototype generator. The experimental results were compared with the corresponding case simulations on the designed generator. The differences between the results predicted by the simulations and those predicted by the measurements were less than 10%. The part and overload performance of the generator has been investigated and it is found from both simulations and measurements that the generator is capable to efficiently operate at varying speeds. The tests on the experimental generator were made for speeds between 2 and 16 rpm and for load variations of 0.5 to 2 per unit. In this thesis it is shown that it is possible to design a very low speed direct drive generator for more or less any given marine current site and this is beneficial for projects aiming to develop a technical and economical viable marine current energy conversion system.

Η ποιότητα πολλών εκ των αποτελεσμάτων της σύγχρονης έρευνας εξαρτώνται άμεσα από την «ποιότητα» και την ποσότητα των τυχαίων αριθμών που χρησιμοποιούνται. Ειδικότερα σε τομείς όπως η στοχαστική μοντελοποίηση και προσομοίωση προτιμώνται οι ντετερμινιστικές γεννήτριες τυχαίων αριθμών, ή αλλιώς γεννήτριες ψευδοτυχαίων αριθμών λόγω της δυνατότητας αναπαραγωγής των αποτελεσμάτων και της μεταφερσιμότητας τους. Επομένως, μας είναι χρήσιμο να εντοπίσουμε ψευδοτυχαίες γεννήτριες αριθμών με αυξημένη φαινόμενη τυχαιότητα αποτελεσμάτων. Για το λόγο αυτό, στη διπλωματική εργασία προτείνεται και εξετάζεται η καταλληλότητα της θραυσματικής διάστασης (fractal dimension) για την αξιολόγηση ψευδοτυχαίων γεννητριών τυχαίων αριθμών (Pseudorandom Number Generators). Η θραυσματική διάσταση αποτελεί μία μετρική που δύναται να εκφράσει την τυχαιότητα των αποτελεσμάτων μιας γεννήτριας ψευδοτυχαίων αριθμών καθώς «ποσοτικοποιεί» την κατανομή των ψευδοτυχαίων αριθμών στον ευκλείδειο χώρο. Σε πρώτο στάδιο γίνεται μία επισκόπηση των υπαρχουσών μεθοδολογιών παραγωγής τυχαίων αριθμών καθώς και των προσεγγίσεων για την αξιολόγηση της απόδοσης των ψευδοτυχαίων γεννητριών τυχαίων αριθμών. Οι καθιερωμένες τεχνικές που εφαρμόζονται για την αξιολόγηση μιας γεννήτριας εστιάζουν σε στατιστικά χαρακτηριστικά που έχουν ως στόχο να μετρήσουν πόσο απρόβλεπτα είναι τα αποτελέσματά της, ή χαρακτηριστικά όπως η περίοδος μιας γεννήτριας. Ακολούθως, μελετάται η θραυσματική διάσταση και οι προτεινόμενες στη βιβλιογραφία μέθοδοι υπολογισμού της. Στο στάδιο αυτό επιλέγεται η κατάλληλη μέθοδος για τον υπολογισμό της θραυσματικής διάστασης. Στο τελευταίο πειραματικό στάδιο παρουσιάζονται τα αποτελέσματα της μέτρησης της μορφοκλασματικής διάστασης. Οι ψευδοτυχαίες γεννήτριες προς αξιολόγηση που μετείχαν στα υπολογιστικά πειράματα ήταν η Γραμμική Αναλογική γεννήτρια, η γεννήτρια Blum-Blum-Shub, η γεννήτρια που βασίζεται στο κρυπτοσύστημα RSA και η γεννήτρια που βασίζεται στο πρόβλημα του διακριτού λογαρίθμου. Τα υπολογιστικά πειράματα επιχειρούν να ανακαλύψουν την απόδοση των εξεταζόμενων γεννητριών αλλά και την ευαισθησία της συμπεριφοράς τους ως προς τις παραμέτρους εισόδου των γεννητριών. / Scientific experimental results are highly dependent on the "quality" and quantity of random numbers used for these experiments. Especially in areas such as stochastic modeling and simulation, deterministic random number generators, known as pseudorandom number generators are preferred because of reproducibility of the results and their portability. Trying to identify pseudorandom number generators sequences which appear to be random, we examine the suitability of Fractal Dimension measurement for assessing Pseudorandom Number Generators. The established techniques that are used to evaluate a generator are focused on statistical features that are designed to detect correlations into generated pseudorandom number sequences. On the other hand, Fractal Dimension is a metric that can express the randomness of the results of a pseudorandom number generator as it "quantifies" the distribution of pseudorandom numbers in Euclidean space. We attempt to evaluate some Pseudorandom Number Generators, like classical Linear Congruential generator, Blum-Blum-Shub generator, the generator based on RSA cryptosystem and the generator based on the Discrete Logarithm problem. The computational experiments presented in our work attempt to assess the performance and the sensitivity of the examined generators.

The thesis presents the design and analysis of a Twin Tensioner for a Belt-driven Integrated Starter-generator (B-ISG) system. The B-ISG is an emerging hybrid transmission closely resembling conventional serpentine belt drives. Models of the B-ISG system's geometric properties and dynamic and static states are derived and simulated. The objective is to reduce the magnitudes of static tension in the belt for the ISG-driving phase. A literature review of hybrid systems, serpentine belt drive modeling and automotive tensioners is included. A parametric study evaluates tensioner parameters with respect to their impact on static tensions. Design variables are selected from these for an optimization study. The optimization uses a genetic algorithm (GA) and a hybrid GA. Results of the optimization indicate the optimal system contains spans with static tensions that are significantly lower in magnitude than that of the original design. Implications of the research on future work are discussed in closing.

Mémoire numérisé par la Division de la gestion de documents et des archives de l'Université de Montréal

The thesis presents the design and analysis of a Twin Tensioner for a Belt-driven Integrated Starter-generator (B-ISG) system. The B-ISG is an emerging hybrid transmission closely resembling conventional serpentine belt drives. Models of the B-ISG system's geometric properties and dynamic and static states are derived and simulated. The objective is to reduce the magnitudes of static tension in the belt for the ISG-driving phase. A literature review of hybrid systems, serpentine belt drive modeling and automotive tensioners is included. A parametric study evaluates tensioner parameters with respect to their impact on static tensions. Design variables are selected from these for an optimization study. The optimization uses a genetic algorithm (GA) and a hybrid GA. Results of the optimization indicate the optimal system contains spans with static tensions that are significantly lower in magnitude than that of the original design. Implications of the research on future work are discussed in closing.

The aim of this study is to present, compare and improve the options of power transmission for tidal current arrays. The potential to generate low or zero carbon power from the world’s tides is increasing as technology moves forward. The technically available tidal current energy resource, the resource that can be captured using existing technology, in the United Kingdom can supply a significant amount of the UK electricity demand. Even though tidal current devices have similarities to offshore wind turbines in many aspects, a number of characteristics differentiate the approach needed regarding power transmission and drive-train design. Some of these characteristics are: predictable direction and speed of the tidal current, predetermined available area in a tidal channel, less swept area due higher density of water, continuous underwater operation and smaller distances to shore. This thesis is based on the hypothesis that tidal current energy can be harnessed using today’s technology in an efficient manner. Technology progression never stops and as new materials and methods become available the cost of utilising tidal current energy will drop in the years to come. However, the research question that has to be asked is whether using today’s technology tidal arrays can be an alternative source of electrical power. In order to respond to this research question electromechanical models of tidal current devices have been developed in detail, from resource to the grid connection, using mathematical linear and non-linear programming in MATLAB/Simulink. The tidal models developed include the tidal resource, the tidal turbine with pitch control, geared induction and synchronous generators, the power electronics with the generator controller, the grid side controller, the cables for power transmission, the filters and the grid connection. All the modelling aspects of this study are presented in Chapter 3. Single tidal current devices were compared using different generator technologies, squirrel cage induction generator or permanent magnet synchronous generator, and different location of the power converters, in the nacelle near the generator or many kilometres apart from the generator. Regarding the generator technology, results showed that even though differences are minor, the permanent magnet synchronous generators are more efficient. Regarding the location of the power converters results showed that positioning the power converters in the nacelle always yields fewer electrical losses but component accessibility is minimised due to the underwater operation of the tidal current device. A key focus aspect of the study is the power transmission option with onshore converters which is presented in detail. Using this concept it is possible to generate electricity from tidal current devices but at the same time keep the highest possible system reliability despite the continuous underwater operation. This concept has been used in the first demonstration tidal current arrays developed by Andritz Hydro Hammerfest. What is more, data provided by Andritz Hydro Hammerfest were utilised in order to validate the simulation models. In this study a step forward is taken regarding the concept of keeping the converter dry and controlling the tidal current generator from afar. An algorithm is developed to design power harmonic filters for systems that use long distance controls. Power harmonic filters allow the long distance control system to operate reliably under all conditions but generate significant electrical losses. The power harmonic filter design algorithm presented in this thesis estimates the exact filter parameters so that the filter ensures maximum system reliability and generate minimum possible losses. In addition tidal array topologies using this concept are developed. The final part of this thesis compares a number of different tidal array topologies based on resource to grid efficiency and component accessibility for maintenance. Results showed that when tidal current devices are clustered per four turbines on offshore platforms it is efficient to use as many clusters as possible connected to a single cable whose both ends are connected to the grid. Locating the power converters in the nacelle yields fewer electrical losses compared to locating the power converters on the offshore platform. However, the difference is minimised because the distance between the tidal current device and the offshore platform is the least possible. Having the power converters on an offshore platform is beneficial in terms of accessibility for maintenance and operation because they are not underwater. The results and the methodology from this thesis can be extended to other offshore renewable energy systems such as the wind and wave. In addition, this study can be used as a stepping stone for decision making by tidal current developers.

Made available in DSpace on 2016-12-12T17:38:33Z (GMT). No. of bitstreams: 1 Andre da Silva Correa.pdf: 3150015 bytes, checksum: 8df17c41e03f4cea1d464cb6c5d5b970 (MD5) Previous issue date: 2015-02-20 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Given the increased contribution of wind farms under complementarity the generating units of electricity derived from petroleum, thermoelectric or hydroelectric energy matrix constituents of countries around the world, this dissertation investigates the advantages and disadvantages between the two main topologies currently used by rogue wind generation. Comparisons between the Doubly Fed Induction Generator and Permanent Magnet Generator are developed in order to establish the different dynamic behavior of each type of wind unit. The constituent parts of wind turbines are studied, performed the modeling and usual control methods are discussed. The simulations were performed in a computer program Matlab / Simulink® whose control models were developed according to the literature, as well as the respective block diagrams for each model. / Diante do crescimento da contribuição dos parques eólicos em regime de complementaridade às unidades geradoras de energia elétrica derivadas de petróleo, termoelétricas ou hidroelétricas constituintes da matriz energética de países espalhados pelo mundo, a presente dissertação de mestrado investiga as vantagens e desvantagens entre as duas principais topologias utilizadas atualmente para geração eólica. São estabelecidas comparações entre o Gerador de Indução Duplamente Alimentado e o Gerador de Ímãs Permanentes, a fim de se estabelecer os diferentes comportamentos dinâmicos de cada tipo de unidade eólica. São estudadas as partes constituintes de aerogeradores, realizadas as modelagens e discutidos os métodos de controle usuais. As simulações foram realizadas em programa computacional Matlab/Simulink®, cujos modelos de controle foram desenvolvidos de acordo com a literatura, assim como os respectivos diagramas de bloco para cada modelo.

This work demonstrates the theoretical and experimental feasibility of a induction generator to work as an attenuator of the harmonics in main generators of on the electric power system, i.e., the synchronous generators. The induction and synchronous machines were analyzed in physical terms and established mathematical modeling in the frequency domain in order to prove the purpose of the thesis. Computer simulations were also carried out in the time domain and laboratory tests were made to check levels of attenuation of current harmonics (THDi) in accordance with the maximum standardized voltage harmonic distortion (THDv). Furthermore, this work presents a study of technical and economical feasibility of a typical application. / Este trabalho demonstra a viabilidade teórica e experimental de um gerador de indução funcionar como um atenuador dos harmônicos nos geradores principais do sistema elétrico de potência, ou seja, os geradores síncronos. Foram analisadas as máquinas de indução e síncronas do ponto de vista físico e estabelecidas modelagens matemáticas no domínio da frequência, a fim de comprovar o objetivo da tese. Foram realizadas também´em, simulações computacionais no domínio do tempo e ensaios em laboratório para verificar os n´níveis de atenuação de harmônicos de corrente (THDi) em conformidade com os limites máximos de distorções harmônicas de tensão (THDv) normalizados. Apresenta-se também neste trabalho um estudo de viabilidade técnica e econômica de uma aplicação típica. / Doutor em Ciências