Detecção de desequilíbrio de tensão em microrredes utilizando redes neurais perceptron
Carregando...
Arquivos
Data
2022-05-02
Autores
Marques, Roberta Carvalho
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
As microrredes podem ser consideradas pequenos conjuntos de cargas e geradores distribuídos, que operam como um único sistema controlável, fornecendo energia e calor para sua área local. Para as concessionárias, as microrredes podem ser consideradas como unidades controladas do sistema de energia. Para o consumidor, as microrredes podem ser projetadas para atender às suas necessidades especiais. Esses sistemas se caracterizam pelo uso extenso de tecnologias da informação, comunicação e automação, permitindo atender a demanda de forma eficiente e otimizada. Os geradores que são utilizados em uma microrrede são geralmente microfontes de energia renovável. Da perspectiva operacional, os geradores devem ser equipados com controles para fornecer a flexibilidade necessária para assegurar que a microrrede opere como um único sistema coletivo e mantenha a Qualidade de Energia Elétrica (QEE) especificadas. Em uma rede de energia elétrica convencional existem geradores síncronos, que são normalmente responsáveis pelo controle de tensão e frequência dos sistemas, o que não se encontra com facilidade em uma microrrede. Quando ocorre o isolamento de uma microrrede, principalmente por causa de eventos não planejados, existe grande probabilidade de ocorrer distúrbios de Qualidade de Energia Elétrica. Com essa exposição e o aumento de instalações de microrredes dá-se a necessidade de metodologias que detectam os problemas de QEE. Considerando a visão das técnicas de detecção de distúrbios de QEE, esta dissertação retrata uma estratégia baseada em Redes Neurais Artificiais perceptron multicamadas em conjunto com simulações do software OpenDSS. Os testes foram realizados em uma microrrede teste adaptada da rede IEEE 13 barras. Os resultados demonstraram-se satisfatórios e efizazes para a detecção dos desequilíbrios de tensão e afundamento e elevação de tensão.
Microgrids are a mix of loads and micro generators that are distributed and operate as a single controllable system, providing energy and heat to its local area. For the dealerships, the microgrid can be considered as a controlled unity of the power system. For the consumer, microgrids can be projected to attend its special needs. The generators that are used in a microgrid are usually micro sources of renewable energy. From the operational perspective, generators must be equipped with controls to provide necessary flexibility to assure that the microgrid operates as a single collective system and keep the specific electrical power quality (EPQ). In a conventional electric power network, there are synchronous generators, that are usually responsible for controlling the tension and frequency of the systems, which is not easily found in a microgrid. When microgrid islanding occurs, mainly because of unplanned events, there is a high probability of voltage imbalance. With this exposed and the increase of microgrids installations, there is a need of methodologies that detect the EPQ problems, especially the voltage imbalance, in the microgrids. Considering the vision of the techniques for detecting EPQ disturbances, this dissertation portrays a methodology based on Artificial Neural Networks, multilayer perceptron combined with OpenDSS software simulation. The tests were performed in a test microgrid adapted from the IEEE 13 bars, with two simulation modes, being one connected to the network (ongrid) and other, disconnected to the general distribution network (offgrid). The results proved to be satisfactory and that they worked effectively for the detection of voltage imbalances.
Microgrids are a mix of loads and micro generators that are distributed and operate as a single controllable system, providing energy and heat to its local area. For the dealerships, the microgrid can be considered as a controlled unity of the power system. For the consumer, microgrids can be projected to attend its special needs. The generators that are used in a microgrid are usually micro sources of renewable energy. From the operational perspective, generators must be equipped with controls to provide necessary flexibility to assure that the microgrid operates as a single collective system and keep the specific electrical power quality (EPQ). In a conventional electric power network, there are synchronous generators, that are usually responsible for controlling the tension and frequency of the systems, which is not easily found in a microgrid. When microgrid islanding occurs, mainly because of unplanned events, there is a high probability of voltage imbalance. With this exposed and the increase of microgrids installations, there is a need of methodologies that detect the EPQ problems, especially the voltage imbalance, in the microgrids. Considering the vision of the techniques for detecting EPQ disturbances, this dissertation portrays a methodology based on Artificial Neural Networks, multilayer perceptron combined with OpenDSS software simulation. The tests were performed in a test microgrid adapted from the IEEE 13 bars, with two simulation modes, being one connected to the network (ongrid) and other, disconnected to the general distribution network (offgrid). The results proved to be satisfactory and that they worked effectively for the detection of voltage imbalances.
Descrição
Palavras-chave
Qualidade de energia elétrica, Microrredes, Redes neurais artificiais, OpenDSS, Voltage imbalance, Electric power quality, Microgrids, Artificial neural networks