Massachusetts Institute of Technology Electronic Systems Laboratory Department of Electrical Engineering. / Bibliography: p. 27-29. / NASA research grant NsG-496 M.I.T. Projects DSR 76152

Massachusetts Institute of Technology Electronic Systems Laboratory Department of Electrical Engineering. / Bibliography: p. 16-20. / NASA research grant NsG-496 M.I.T. Projects DSR 76152

Made available in DSpace on 2014-06-11T19:30:51Z (GMT). No. of bitstreams: 0 Previous issue date: 2013-04-26Bitstream added on 2014-06-13T19:40:20Z : No. of bitstreams: 1 silva_ef_dr_ilha.pdf: 1154439 bytes, checksum: 6d38d1a30422651449a03143bab1be4d (MD5) / Neste trabalho é apresentado um modelo estocástico linear inteiro misto para o problema de planejamento da expansão da rede de transmissão multiestágio considerando restrições de segurança N −1 (PERTMRS) a longo prazo. Considerando uma amostragem de cenários para a demanda e geração (com uma probabilidade para cada cenário) é possível transformar o modelo estocástico proposto num equivalente determinístico linear inteiro misto (LIM). O uso de um modelo LIM garante a convergência para a solução ótima do PERTMRS usando métodos de otimização clássica existentes. O critério de segurança N −1 indica que o sistema de transmissão deve ser expandido de tal forma que, com a saída de operação de uma linha existente ou candidata (em um conjunto pré-definido de contingências) do sistema, o mesmo ainda deve operar adequadamente. O modelo foi implementado usando a linguagem de modelagem algébrico AMPL e solucionado usando o solver comercial CPLEX. Os sistemas de testes: Garver de 6 barras e IEEE de 24 barras; e os sistemas reais: Colombiano de 93 barras e o Boliviano de 57 barras foram usados para avaliar o modelo proposto. Para os sistemas de grande porte uma estrategia de redução do espaço de busca combinatório do problema é apresentado para facilitar a implementação do modelo. / In this work we present a mixed integer linear stochastic model for the long term multistage transmission expansion planning problem considering N −1 security constraints (PERTMRS). Considering a sampling for each demand and generation scenario (with a predefined probability for each scenario), the proposed stochastic model can be transformed to a deterministic mixed integer linear programming problem (LIM). The use o LIM model gurantess the convergence to the optimum solution of the PERTMRS if a classical optimization techniques is employed. The N −1 safety criterion indicates that the transmission system must be expanded such that, with an outage of an existing or candidate line (from a predefined set of contingencies, the system should still operate properly. The model was implemented using the algebraic modeling language AMPL and solved using the commercial solver CPLEX. The 6-bus Garver and the IEEE-24 buses test systems and the real 93-bus Colombian and 57-bus Bolivian systems were used to evaluate the proposed model. For large systems a strategy to reduce the combinatorial search space of the problem is presented to facilitate implementation of the model.

Made available in DSpace on 2014-06-11T19:22:33Z (GMT). No. of bitstreams: 0 Previous issue date: 2013-06-24Bitstream added on 2014-06-13T20:29:05Z : No. of bitstreams: 1 garciaosorio_va_me_ilha.pdf: 7086249 bytes, checksum: f0c9e3175a083de7a457f0924775c2ef (MD5) / O escopo deste trabalho é a otimização do carregamento dos Veículos elétricos (VEs) em um ambiente controlado, considerando-se uma infraestrutura de medição avançada. O método desenvolvido utiliza uma técnica de otimização linear, a qual determina a potência elétrica que será entregue a cada VE considerando como restrições os níveis de operação da rede elétrica de média tensão. Essa potência será chamada neste trabalho como taxa de carregamento ótimo. Para a aplicação da metodologia proposta neste trabalho, é necessário conhecer a distribuição espacial dos veículos elétricos híbridos plug-in (PHEVs) na cidade e a distância percorrida por cada um deles. Para isto, foi utilizado o modelo espaço-temporal que utiliza um sistema multiagente e um enfoque de percolação o qual permite estimar a localização de cada PHEV e o estado de carga inicial (SOC) para a conexão no sistema. Estes dados são considerados como dados de entrada para a metodologia apresentada neste trabalho. O método foi testado no sistema teste IEEE 34 considerando niveis de penetração de 15%, 30% e 50% e o algoritmo desenvolvido foi implementado em linguagem de programação MATLAB 7.9.0 (R2009a). Os resultados mostram que com uma porcentagem de penetração de 15%, é possível atender a demanda adicional, representada pela conexão dos PHEVs, sem violar as restrições de operação do sistema; com uma porcentagem de penetração de 30%, os resultados mostram que, preservando os limites de operação do sistema, uma porcentagem de 20,66% dos PHEVs consegue carregar totalmente suas baterias. Também, com uma porcentagem de penetração de 50% os resultados mostram que 8,56% dos PHEVs consegue ficar dentro de um nível de energia aceitável. Porém, gerando cenários de recarga onde os níveis de tensão do sistema são aumentados e... / The scope of this work is to optimize the charging of electric vehicles (EVs) in a controlled environment, considering an advanced metering infrastructure. The method uses a linear optimization technique, which determines the electrical power that will be delivered to each EV considering the levels of operating constraints in a medium-voltage distribution network. In this work, such power is called the optimal charging rate. For the application of the methodology proposed in this work, it is necessary to know the spatial distribution of the hybrid electric vehicles (PHEVs) and the distance traveled by each of them. For this, this work used the spatial-temporal model formed by a multi-agent system with a percolation approach to estimate the location of each PHEV, and the initial state of charge (SOC) in the beginning of recharge. These data are considered as inputs of the presented methodology in this work. The method was tested in the IEEE 34 test system considering penetration levels of 15%, 30% and 50% and the developed algorithm was implemented in the MATLAB programming language 7.9.0 (R2009a). The results show that with 15% of penetration, it is possible to meet the additional demand, represented by the connection of PHEVs without violating the operation system constraints; with a percentage of 30% penetration, the results show that while preserving the operating limits of the system, a percentage of 20.66% of PHEVs can fully charge their batteries. Also, with a percentage of 50% penetration results show that 8.56% of PHEVs can stay within an acceptable power level. However, creating scenarios where the system voltage levels are increased and, moreover, the charging time is extended, the batteries of all PHEVs, representing a rate of penetration of 30% are in the end of the clearance within an acceptable level and, for 50%, a... (Complete abstract click electronic access below)

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 / This thesis focuses on the design, development and real-time simulation of a robust nonlinear networked control for the dissolved oxygen concentration as part of the wastewater distributed systems. This concept differs from previous methods of wastewater control in the sense that the controller and the wastewater treatment plants are separated by a wide geographical distance and exchange data through a communication medium. The communication network introduced between the controller and the DO process creates imperfections during its operation, as time delays which are an object of investigation in the thesis. Due to the communication network imperfections, new control strategies that take cognisance of the network imperfections in the process of the controller design are needed to provide adequate robustness for the DO process control system. This thesis first investigates the effects of constant and random network induced time delays and the effects of controller parameters on the DO process behaviour with a view to using the obtained information to design an appropriate controller for the networked closed loop system. On the basis of the above information, a Smith predictor delay compensation controller is developed in the thesis to eliminate the deadtime, provide robustness and improve the performance of the DO process. Two approaches are adopted in the design of the Smith predictor compensation scheme. The first is the transfer function approach that allows a linearized model of the DO process to be described in the frequency domain. The second one is the nonlinear linearising approach in the time domain. Simulation results reveal that the developed Smith predictor controllers out-performed the nonlinear linearising controller designed for the DO process without time delays by compensating for the network imperfections and maintaining the DO concentration within a desired acceptable level. The transfer function approach of designing the Smith predictor is found to perform better under small time delays but the performance deteriorates under large time delays and disturbances. It is also found to respond faster than the nonlinear approach. The nonlinear feedback linearisig approach is slower in response time but out-performs the transfer function approach in providing robustness and performance for the DO process under large time delays and disturbances. The developed Smith predictor compensation schemes were later simulated in a real-time platform using LabVIEW. The Smith predictor controllers developed in this thesis can be applied to other process control plants apart from the wastewater plants, where distributed control is required. It can also be applied in the nuclear reactor plants where remote control is required in hazardous conditions. The developed LabVIEW real-time simulation environment would be a valuable tool for researchers and students in the field of control system engineering. Lastly, this thesis would form the basis for further research in the field of distributed wastewater control.

M.Tech. (Electrical and Electronic Engineering) / With the increasing use of nonlinear loads in power systems, harmonic pollution becomes more frequent with disastrous consequences on the electrical network and other loads. Harmonic currents cause problems such as system resonance, protective relays and switchgear malfunction, communication interference, incorrect meter, equipment and process failure, over heating of transmission and distribution systems, insulation degradation and power transformer failure. It appears therefore absolutely essential to determine the current profile of the loads connected to the same Point of Common Coupling (PCC). This study aims to assess a distorted power system in order to determine the characteristics components of distorted waveforms (DC components, fundamental and harmonics components), and to quantify the respective distortion contributions in the power network. The numerical method is applied to analyse the current waveforms captured with the aid of the digital scope meter (DSO) 1200 series at different sections of the network. The power network under study is designed, constructed and analysed with the MatLab/Simulink 7.0 package. In both methods, the Individual Harmonic Distortion (IHD) and the current total harmonic distortion (ITHD) for the line, the non-linear and linear loads are computed for the harmonic components 5th, 7th, 11th and 13th. The results showed that for the analysis of the power system, numerical analysis based on the resultant harmonic curve method and the MatLab/Simulink as applied to the network all converged to highlight that harmonic components appear both on the line, the PCC and the linear load. This is due to the fact that the converter characteristic harmonics (5th, 7th, 11th, 13th…) are found in the network components (line, PCC, linear load…). The resultant harmonic curve method indicated a randomly distributed percentage IHD in different sections of the network for harmonic orders higher than 7th, whereas the Matlab/Simulink results normally show a non-uniform decay pattern of the IHD as the harmonic order increases within the respective section of the network.

Orientador: Antonio Padilha Feltrin / Coorientador: Augusto César Rueda Medina / Banca: Fábio Bertequini Leão / Banca: Gelson Antonio Andrêa Brigatto / Resumo: O escopo deste trabalho é a otimização do carregamento dos Veículos elétricos (VEs) em um ambiente controlado, considerando-se uma infraestrutura de medição avançada. O método desenvolvido utiliza uma técnica de otimização linear, a qual determina a potência elétrica que será entregue a cada VE considerando como restrições os níveis de operação da rede elétrica de média tensão. Essa potência será chamada neste trabalho como taxa de carregamento ótimo. Para a aplicação da metodologia proposta neste trabalho, é necessário conhecer a distribuição espacial dos veículos elétricos híbridos plug-in (PHEVs) na cidade e a distância percorrida por cada um deles. Para isto, foi utilizado o modelo espaço-temporal que utiliza um sistema multiagente e um enfoque de percolação o qual permite estimar a localização de cada PHEV e o estado de carga inicial (SOC) para a conexão no sistema. Estes dados são considerados como dados de entrada para a metodologia apresentada neste trabalho. O método foi testado no sistema teste IEEE 34 considerando niveis de penetração de 15%, 30% e 50% e o algoritmo desenvolvido foi implementado em linguagem de programação MATLAB 7.9.0 (R2009a). Os resultados mostram que com uma porcentagem de penetração de 15%, é possível atender a demanda adicional, representada pela conexão dos PHEVs, sem violar as restrições de operação do sistema; com uma porcentagem de penetração de 30%, os resultados mostram que, preservando os limites de operação do sistema, uma porcentagem de 20,66% dos PHEVs consegue carregar totalmente suas baterias. Também, com uma porcentagem de penetração de 50% os resultados mostram que 8,56% dos PHEVs consegue ficar dentro de um nível de energia aceitável. Porém, gerando cenários de recarga onde os níveis de tensão do sistema são aumentados e... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: The scope of this work is to optimize the charging of electric vehicles (EVs) in a controlled environment, considering an advanced metering infrastructure. The method uses a linear optimization technique, which determines the electrical power that will be delivered to each EV considering the levels of operating constraints in a medium-voltage distribution network. In this work, such power is called the optimal charging rate. For the application of the methodology proposed in this work, it is necessary to know the spatial distribution of the hybrid electric vehicles (PHEVs) and the distance traveled by each of them. For this, this work used the spatial-temporal model formed by a multi-agent system with a percolation approach to estimate the location of each PHEV, and the initial state of charge (SOC) in the beginning of recharge. These data are considered as inputs of the presented methodology in this work. The method was tested in the IEEE 34 test system considering penetration levels of 15%, 30% and 50% and the developed algorithm was implemented in the MATLAB programming language 7.9.0 (R2009a). The results show that with 15% of penetration, it is possible to meet the additional demand, represented by the connection of PHEVs without violating the operation system constraints; with a percentage of 30% penetration, the results show that while preserving the operating limits of the system, a percentage of 20.66% of PHEVs can fully charge their batteries. Also, with a percentage of 50% penetration results show that 8.56% of PHEVs can stay within an acceptable power level. However, creating scenarios where the system voltage levels are increased and, moreover, the charging time is extended, the batteries of all PHEVs, representing a rate of penetration of 30% are in the end of the clearance within an acceptable level and, for 50%, a... (Complete abstract click electronic access below) / Mestre

In this work passive piezoelectric devices for vibration damping are studied. It is developed the basic idea of synthesizing electrical wave guides to obtain an optimal electro-mechanical energy exchange and therefore to dissipate the mechanical vibrational energy in the electric form. Modular PiezoElectroMechanical (PEM) structures are constituted by continuous elastic beams (or bars) coupled, by means of an array of PZT transducers, to lumped dissipative electric networks. Both refined and homogenized models of those periodic systems are derived by an energetic approach based on the principle of virtual powers. Weak and strong formulation of the dynamical problem are presented having in mind future studies involving the determination of numerical solutions. In this framework the effectiveness of the proposed devices for the suppression of mechanical vibrations is investigated by a wave approach, considering both the extensional and flexural oscillations. The optimal values of the electric parameters for a fixed network topology are derived analytically by a pole placement technique. Their sensitivities on the dimensions of the basic cell of the periodic system and on the design frequency are studied. Moreover the dependence of damping performances on the frequency is analyzed. Comparing the performances of different network topological configurations, the advantages of controlling a mechanical structure with an electric analog are shown. As a consequence of those results, new interconnections of PZT transducers are proposed. An experimental setup for the validation of the analytical and numerical results is proposed and tested. A classical experience on resonant shunted PZT is reproduced. Future experimental work is programmed. / Master of Science

Orientador: José Roberto Sanches Mantovani / Banca: Ruben Augusto Romero Lázaro / Banca: Benemar Alencar de Souza / Resumo: Neste trabalho apresentam-se o desenvolvimento e a implementação de um algoritmo computacional para o problema de planejamento e projeto de circuitos secundários de distribuição de energia elétrica. O problema é formulado como um modelo de Programação Não Linear Inteiro Misto (PNLIM), e para sua solução propõe-se um algoritmo de Busca Tabu (BT). Esta ferramenta permite determinar e avaliar os custos dos elementos que compõem o sistema de distribuição (reforma e/ou construção de redes secundárias), evidenciando sua robustez e precisão. Os algoritmos de cálculo mecânico e definição de estruturas são baseados em normas técnicas para se determinar os postes e estruturas do sistema sob estudo. Os condutores são especificados pelo algoritmo de fluxo de potência trifásico respeitando as condições topológicas, físicas e ambientais. Toda metodologia de cálculo e projeto utilizada neste problema está de acordo com os Procedimentos de Distribuição do Sistema Elétrico Nacional (PRODIST). Para mostrar a eficiência do modelo proposto e da metodologia de planejamento e projeto desenvolvidos, apresentam-se resultados para dois sistemas reais de distribuição com 76 e 54 barras, respectivamente. / Abstract: In this work, the development and implementation of a computational algorithm for the planning and design of secondary circuits of power distribution systems problem, is presented. The problem is formulated as a Non-linear Integer Mixed Programming (NLIMP) problem; and for its solutions it is proposed a Tabu Search (TS) algorithm. This tool allows determining and evaluating the cost of elements that are part of the distribution system (upgrade and/or construction of secondary networks), showing its robustness and accuracy. Mechanic calculation algorithms and structure definitions are based on technical norms in order to determine poles and structures of the system under study. Conductors are specified by the three-phase power flow algorithm, satisfying topological, physical and environmental conditions. All the calculation and design methodology used in this problem agrees with the Distribution Procedures of the National Electric System (PRODIST). In order to show the efficiency of the proposed model and the developed planning and design methodology, results for to real-life distribution systems, with 76 and 54 buses, are presented. / Mestre

The automation of power distribution networks contributes to the improvement of the quality indicators for providing isolation of defective parts through remote controls for opening and closing devices, as well as transfers to defective parts without alternative sources of supply. These commands are normally executed in the Operation Centers by specialized technicians. Although, providing greater speed than the local execution of opening and closing braces through teams in the field for more complex circuits requires technicians operating an analysis of a significant set of variables for decision making. This time devoted to technical analysis reduces earnings quality compared automation solution, where commands could be achieved through pre-defined algorithms, known as Automatic Replenishment. To measure the gain of additional solutions of this nature is necessary to determine the behavior of time to maneuver performed by technicians and those obtained by automatic recomposition. With such a statistical study sample can define the average times for each situation and size of these times you can compare the two situations, basically inserting the time difference on a methodology for calculation of quality indicators. This study aims to perform the statistical survey on a sample of a system already deployed in a power distributor, to establish comparison regarding the quality gains through the methodology of the MLE (Logical Structural) and apply a decision model based on multicriteria to provide a prioritization matrix circuits where automatic recovery can be more advantageous. Although the scope of this work involves only the automatic replenishment, the statistical basis of the obtained times can serve as a starting point for studies related to other more sophisticated mechanisms of self-diagnosis and self restoration as the "Self Healing" present in the context of Smart Grids. / A automação das redes de distribuição de energia contribui para melhoria dos indicadores de qualidade por proporcionar o isolamento dos trechos defeituosos através de comandos a distância de abertura e fechamento dos dispositivos, além de transferências de trechos sem defeito para fontes alternativas de suprimento. Estes comandos normalmente são executados nos Centros de Operação por técnicos especializados. Embora, proporcione maior celeridade do que a execução local de abertura e fechamento de chaves através de equipes em campo, para circuitos mais complexos exige dos técnicos de operação uma análise de um conjunto significativo de variáveis para tomada de decisão. Este tempo dedicado de análise dos técnicos reduz os ganhos de qualidade da automação em comparação a soluções de recomposição automática, onde os comandos pudessem ser realizados através de algoritmos pré-definidos. Para aferir o ganho adicional de soluções dessa natureza é necessário apurar o comportamento dos tempos de manobra executados pelos técnicos e daqueles obtidos pela recomposição automática. Com tal amostra um estudo estatístico pode definir os tempos médios para cada situação e de porte destes tempos pode-se comparar as duas situações, basicamente inserindo a diferença de tempos em uma metodologia para cálculo de indicadores de qualidade. Este trabalho tem o objetivo de executar este estudo estatístico em uma amostra de um sistema já implantado em uma distribuidora de energia, estabelecer comparação referente aos ganhos de qualidade através da metodologia da MLE (Matriz Lógica Estrutural) e aplicar um modelo de decisão baseado em multicritério para fornecer uma matriz de priorização de circuitos onde a recomposição automática possa ser mais vantajosa. Esta base estatística obtida dos tempos pode servir como ponto de partida para outros estudos relacionados a mecanismos mais sofisticados de autodiagnostico e de auto restabelecimento como o Self Healing presente no contexto das redes inteligentes ou Smart Grids .