Estimação robusta de modelo funcional tridimensional (3D), para identificação de faltas em linhas de transmissão

• Ano de defesa: 2016
• Autor(a) principal: Gilmar Rosa
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; Brasil; ENG - DEPARTAMENTO DE ENGENHARIA PRODUÇÃO; Programa de Pós-Graduação em Engenharia de Produção; UFMG
• Palavras-chave em Português: Linhas de transmissão; Análise funcional; Regressão quantílica; Classificadores; Wavelet; Engenharia de produção; Linhas de telecomunicação
• Link de acesso: http://hdl.handle.net/1843/41713

This thesis presents a novel method for detection and fault classification 2D and its extension to 3D, that is, taking into account the three phases of the transmission lines (LT) based on the functional form of the ellipse. the methodology implementation of Algorithms are presented, as well as tests in real and simulated data. The results were satisfactory indices with similar accuracy to those found in the literature which uses the wavelet transform pattern. It is proposed construction elliptical confidence limits through quantile regression, which are used for monitoring the transmission line to the fault detection purpose. Thus upon detection of the fault characteristics of the signals are extracted faulted via distance metric as input and used in binders in order to determine the type of fault that has occurred. Thus, the simplicity of the methodology and implementation are intelligible inherent characteristics, which can make your most extensive use.