Sistema automático de detecção de falhas em rolamentos de roletes de correias transportadoras

Detalhes bibliográficos
Ano de defesa: 2010
Autor(a) principal: Marina Machado Loures
Orientador(a): Não Informado pela instituição
Banca de defesa: Não Informado pela instituição
Tipo de documento: Dissertação
Tipo de acesso: Acesso aberto
Idioma: por
Instituição de defesa: Universidade Federal de Minas Gerais
UFMG
Programa de Pós-Graduação: Não Informado pela instituição
Departamento: Não Informado pela instituição
País: Não Informado pela instituição
Palavras-chave em Português:
Link de acesso: http://hdl.handle.net/1843/BUOS-8D2FWU
Resumo: The rollers are the main elements responsible for the movement of conveyor belts, and its integrity is fundamental to the appropriate operation of the conveyors. To increase the operational reliability and to reduce the need for preventive maintenance, an efcient system process monitoring is necessary, which ideally should allow the detection of incipient failures. The objective of this work is the detection of faults in the internal bearings of rollers, after all they are responsible for most of the failures and lockings on rollers. In addition, bearing failure detection techniques are widely found in literature. Rolling bearings vibration signals have been used to fault detection and classication by means ofcareful analysis of the corresponding signals spectral content. A usual procedure consists of searching signicant peaks in the signal power spectral density (PSD) in frequency bands corresponding to specic types of faults. However, this technique sometimes is not reliable once there could be a high degree of subjectivism in the evaluation of lowsignal-to-noise ratio signals, or when an inadequate windowing is applied to the original time series. In both cases, one has the production of artifacts as spurious peaks in the frequency domain. In order to minimize this problem, the information on the probability that a given numerically determined PSD local maximum is indeed a peak is added to a fuzzy inference machine. Moreover, the importance of this maximum PSD value is also evaluated by comparing it with the white noise PSD estimated value. Preliminary results have shown that this procedure can be used to enhance the specialist knowledge representation in a fuzzy system used to perform rolling bearing fault diagnosis.