O diagnóstico de transitórios oscilatórios em sistemas de distribuição de média tensão pelo uso da morfologia matemática e da busca por correspondência

• Ano de defesa: 2021
• Autor(a) principal: Hoffmann, Alessandra Bortoletto Garbelotti
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade de São Paulo; Medianeira; Brasil; Programa de Pós-Graduação em Engenharia Elétrica; USP
• Palavras-chave em Português: Transitórios (Eletricidade); Energia elétrica – Distribuição; Energia elétrica – Transmissão; Transients (Electricity); Electric power distribution; Electric power transmission; CNPQ::ENGENHARIAS::ENGENHARIA ELETRICA; ENGENHARIA ELÉTRICA
• Link de acesso: http://repositorio.utfpr.edu.br/jspui/handle/1/29184

The objective of this research, in the context of Power Quality (PQ), is the diagnosis and the classification of Oscillatory Transients (OTs) caused by switching specific loads in electric power systems. Although the transient occurs in a short period, it is of great importance due to its duration, amplitude, and manifested frequencies, being possible to observe it via the three-phase voltages and/or currents registered in the Medium Voltage Distribution Systems (MVDSs). The switches considered as sources for the OTs were: an energized line and capacitor bank. A comparative analysis of the processing time and the mean error in detection, using the techniques Notch Filter, Wavelet Transform, and Mathematical Morphology (MM) to identify the beginning of OTs in synthetic signals was performed, with and without the presence of noise. Using analytic equations, a synthetic database of OTs was created, thus forming a dictionary used to extract the parameters. For the detection of signal discontinuity, the MM technique presented better results, and with the application of a matching search algorithm, or Matching Pursuit Algorithm, it were possible to classify the source causing the OTs. To evaluate the proposed approach, the OTs were represented from distinct and multiple synthetic and random signals, as well as experimental real signals acquired from the IEEE working group (Working Group P1159.3). The results obtained showed that it is possible to classify the specific sources that generate the OTs considering an individual event, as well as multiple events in a MVDS.