Detecção de barras falhadas em motores de indução utilizando um filtro gaussiano passa faixa

Three-phase induction motors are probably today the most important rotary machines in the industrial sector since they are widely used in the most varied applications and are present in almost all types of drives.It is estimated that almost 80% of the total electric motors used in the industry for t...

ver descrição completa

Detalhes bibliográficos
Autor: Souza, Mateus Ventura
Tipo de documento: dissertação
Estado:Versão publicada
Data de publicação:2018
País:Brasil
Recursos:Universidade Federal de Sergipe (UFS)
Repositório:Repositório Institucional da UFS
Idioma:português
OAI Identifier:oai:oai:ri.ufs.br:repo_01:riufs/10074
Acesso em linha:http://ri.ufs.br/jspui/handle/riufs/10074
Access Level:Acceso aberto
Palavra-chave:Engenharia elétrica
Motores elétricos de indução
Filtros elétricos
Filtros elétricos digitais
Motor de indução
Filtros
Filtros digitais
Filtro passa faixa
Análise de falhas
Diagnóstico de falhas
Barras falhadas
Induction motor
Filters
Digital filters
Band pass filter
Failure analysis
Fault diagnostics
Failed bars
ENGENHARIAS::ENGENHARIA ELETRICA
Descrição
Resumo:Three-phase induction motors are probably today the most important rotary machines in the industrial sector since they are widely used in the most varied applications and are present in almost all types of drives.It is estimated that almost 80% of the total electric motors used in the industry for transformation from electrical energy to mechanical energy are Induction Engines. Of course this provides the search for tools that offer greater efficiency and safety to engine maintenance systems. A problem of fundamental importance in the predictive maintenance process of these electric motors is the detection of failures in advance to avoid the production stoppage. This work proposes to detect faults in the rotor bars. In the state of the state there are several techniques covered, some using the engine in its steady state and others using the engine starting. In all of them the detection of the fault in the bar occurs, at its best, after the complete failure of the first bar. This work, then, proposes a new methodology for detecting the failure of the bar when it has not yet completely failed, but with increased resistance in the failed bar circuit. For this, a Gaussian bandpass filter with optimized selection of bandwidth and frequency bandwidth was used.This filter acts as an optimization of the Discrete Wavelet Transform and detects the presence of the fault in the bar using the motor start signal. The proposed technique was applied in motors with rotors manufactured by the company WEG specifically for this work and with previously known characteristic faults and was also applied in simulated rotors. The quantification of the degree of fault severity was done using the Teager Energy tool. The experiments were done through simulations in a mathematical simulation software and in real engines with bench tests and the results proved the effectiveness of the proposed method.