Mapeamento da vegetação para análise de faltas no sistema de distribuição do município de Presidente Prudente-SP

Carregando...
Imagem de Miniatura

Data

2024-07-10

Orientador

Boschi, Letícia Sabo

Coorientador

Pós-graduação

Curso de graduação

Rosana - FEC - Engenharia de Energia

Título da Revista

ISSN da Revista

Título de Volume

Editor

Universidade Estadual Paulista (Unesp)

Tipo

Trabalho de conclusão de curso

Direito de acesso

Acesso abertoAcesso Aberto

Resumo

Resumo (português)

As interrupções de energia em redes de distribuição de alta, média e baixa tensão, causadas pela vegetação arbórea são um grande problema no sistema de redes aéreas, o mais utilizado no Brasil. Visto essa problemática, um caminho para a solução deste tópico é o mapeamento da vegetação, utilizando classificadores de alta precisão. O mapeamento da vegetação, quando efetuado de maneira assertiva, fornece dados qualitativos e quantitativos para tomadas de decisões efetivas e preventivas, como por exemplo as podas de árvores com maiores riscos de interrupção na rede. Esse estudo foi realizado a partir das ferramentas QGIS e Google Earth Engine (GEE). Utilizando a ferramenta GEE, foi possível obter imagens do satélite Sentinel 2A, e classificá-las em 6 classes distintas. A precisão dos resultados se tornou possível por conta do algoritmo de classificação utilizado, o Random Forest, e os índices espectrais NDVI, MNDWI e NDBI. A ferramenta GEE se mostrou eficiente para o objetivo deste trabalho, uma vez que aplicado os índices de qualidade, Kappa e Acurácia Global (AC), estes retornaram com precisão acima de 91% em todas as classes de classificação. Após obter a imagem com classificação precisa, foi feito o tratamento de dados na ferramenta QGis, uma análise dos dados de redes de distribuição que passam no município de Presidente Prudente, obtidos através do portal da ANEEL (Agência Nacional de Energia Elétrica). Com os dados das linhas de baixa, média e alta tensão, foi possível sobrepor a classificação da vegetação arbórea e indicar quais áreas têm maior risco de interrupções na rede. Após realizar essa análise, foi efetuado um levantamento com dados quantitativos das interrupções por Fator Gerador (FG), constatando que a vegetação arbórea é o quarto maior FG dentre os apresentados no relatório da ANEEL. A partir dos resultados apresentados, é possível afirmar que o mapeamento da vegetação permite que ações preventivas sejam executadas a fim de evitar interrupções na rede, como podas de árvores, uma vez que as localidades com maiores riscos são identificadas.

Resumo (inglês)

Power interruptions in high, medium and low voltage distribution networks, caused by tree vegetation, are a major problem in the aerial network system, the most used in Brazil. Given this problem, one way to resolve this topic is vegetation mapping, using high-precision classifiers. Vegetation mapping, when carried out assertively, provides qualitative and quantitative data for effective and preventive decision-making, such as pruning trees with greater risks of interruption to the network. This study was carried out using QGIS and Google Earth Engine (GEE) tools. Using the GEE tool, it was possible to obtain images from the Sentinel 2A satellite and classify them into 6 different classes. The precision of the results became possible due to the classification algorithm used, Random Forest, and the spectral indices NDVI, MNDWI and NDBI. The GEE tool proved to be efficient for the objective of this work, since when applying the veracity, Kappa and Global Accuracy (AC) indices, these returned with accuracy above 91% in all classification classes. After obtaining the image with precise classification, data processing was carried out using the QGis tool, an analysis of data from distribution networks that pass through the municipality of Presidente Prudente, obtained through the ANEEL (National Electric Energy Agency) portal. With data from low, medium and high voltage lines, it was possible to overlay the classification of tree vegetation and indicate which areas are at greater risk of network interruptions. After carrying out this analysis, a survey was carried out with quantitative data on interruptions by Generating Factor (FG), finding that tree vegetation is the fourth largest FG among those presented in the ANEEL report. From the results presented, it is possible to affirm that vegetation mapping allows preventive actions to be carried out in order to avoid interruptions in the network, such as tree pruning, since the locations with the greatest risks are identified.

Descrição

Idioma

Português

Como citar

Itens relacionados

Financiadores