Transformada wavelet aplicada a análise automática de oscilografias de curta duração em unidades geradoras

• Ano de defesa: 2016
• Autor(a) principal: Pereira, Sidnei
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Tecnológica Federal do Paraná; Pato Branco; Brasil; Programa de Pós-Graduação em Engenharia Elétrica; UTFPR
• Palavras-chave em Português: Sistemas de energia elétrica; Osciladores elétricos; Wavelets (Matemática); Transitórios (Eletricidade); Electric power systems; Oscillators, Electric; Wavelets (Mathematics); Transients (Electricity); CNPQ::ENGENHARIAS::ENGENHARIA ELETRICA
• Link de acesso: http://repositorio.utfpr.edu.br/jspui/handle/1/1759

The focus of this work is the automatic analysis of disturbance records for electrical power generating units. The main proposition is a method based on wavelet transform applied to short-term disturbance records (waveform records). The goal of the method is to detect the time instants of recorded disturbances and extract meaningful information that characterize the faults. The result is a set of representative information of the monitored signals in power generators. This information can be further classified by an expert system (or other classification method) in order to classify the faults and other abnormal operating conditions. The large amount of data produced by digital fault recorders during faults justify the research of methods to assist the analysts in their task of analysing the disturbances. The literature review pointed out the state of the art and possible applications for oscillography records. The review of the COMTRADE standard and wavelet transform underlines the choice of the method for solving the problem. The conducted tests lead to the determination of the best mother wavelet for the segmentation process. The application of the proposed method to five case studies with real oscillographic records confirmed the accuracy and efficiency of the proposed scheme. With this research, the post-operation analysis of occurrences is improved and as a direct result is the reduction of the time that generators are offline.