Detecção e classificação de faltas de curto-circuito em sistemas de distribuição com a inserção de geração distribuída

Carregando...
Imagem de Miniatura

Data

2020-12-14

Autores

Faria, Andréia da Silva Santos

Título da Revista

ISSN da Revista

Título de Volume

Editor

Universidade Estadual Paulista (Unesp)

Resumo

A detecção e classificação de faltas de curtos-circuitos são de fundamental importância para minorar danos a equipamentos elétricos e prevenir faltas de regime permanente ou blecautes em sistemas de Distribuição. A inserção de geração distribuída na rede de distribuição altera a topologia do sistema elétrico e afeta diretamente a corrente de curto-circuito. Nesse contexto, neste trabalho propõe-se uma metodologia alternativa para detecção e classificação de faltas de curto-circuito em sistemas de distribuição de energia elétrica trifásicos considerando a inserção de geração distribuída. A detecção baseia-se nos índices comportamentais das correntes de curto-circuito trifásicas que são extraídos por meio da Transformada Wavelet Discreta de sobreposição Máxima e Análise de Multirresolução. A classificação das fases em curto-circuito é realizada via uso da rede neural ARTMAP-Fuzzy. As diversas condições de falta foram simuladas no sistema teste IEEE-34 barras modificado. O sistema teste foi modelado no software ATPDraw e a metodologia para detecção e classificação de faltas foi implementada no software MATLAB. A técnica para detecção e classificação é robusta. Foram detectados corretamente em todos os cenários implementados, mais de 98% dos curtos-circuitos, enquanto o acerto do classificador foi superior a 98% com apenas uma execução ou época para as etapas de treinamento e teste, portanto o mesmo apresenta baixo esforço computacional.
The detection, as well as the classification of faults relative to short-circuit, are of fundamental importance to mitigate damages to electrical equipment and prevent faults of permanent regime or blackouts in electrical power systems. The insertion of distributed generation in the distribution network changes the topology of the electrical system and directly affects the shortcircuit current. In this context, an alternative metholody aimed at detecting and classifying short-circuit faults in three-phase power distribution systems is proposed in this work with the insertion of distributed generation. The detection is made from the evaluation of the indices of three-phase short-circuit currents. These indices are obtained through the Discrete Wavelet Transform and the Multi-Resolution Analysis. The classification of the short-circuit phases is performed using the ARTMAP-Fuzzy neural network. The various fault conditions were simulated in the modified IEEE 34-bus. That test system was represented in the ATPDraw software and the methodology for fault detection and classification was implemented in the MATLAB software. The detection and classification technique is robust. More than 98% of short circuits were correctly detected, while the classifier reached 98% assertiveness with just one training and test period with low computational effort.

Descrição

Palavras-chave

Análise multirressolução, Detecção de faltas de curto-circuito, Rede neural ARTMAP Fuzzy, Sistemas de distribuição de energia elétrica, Transformada Wavelet discreta de sobreposição máxima, Electrical energy distribution systems, Multi-resolution analysis, Maximal overlap discrete Wavelet transform, Neural network Fuzzy ARTMAP, Short-circuit fault detection

Como citar