SOUZA, B. A.; http://lattes.cnpq.br/8757457757127097; SOUZA, Bruno Almeida de.
Résumé:
The objective of this work is to propose a polymeric insulators classification technique
that allows the recommendation of its state of degradation in operation, in order to
determine the most appropriate time to perform maintenance (exchanges), thus
contributing to the reduction of interruptions . The technique relies on analysis of the
infrared radiation (IF) emitted by the tested insulators. Polymeric insulators 230 kV
were used with varying levels of degradation as test objects. The IF images obtained
during the test were subjected to image processing using the RGB type image in order
to minimize the existing noise. A neural network to classify the state of degradation of
polymeric insulators was proposed. The network entries are the variations in
temperature in the body of the insulator and temperature obtained from the
measurement of the RIF. The network output was the degree of practicality the state of
degradation of the insulators. The results show the technical efficiency in aid to
decision-making, the need to replace or not the insulators. It is emphasized that the
isolation technique of measuring infrared radiation, thermal imaging is not sufficient for
the diagnosis of insulators, however it is essential to diagnose faults on polymeric
insulators.
