Classificação de isoladores de vidro de alta tensão a partir do espectro eletromagnético irradiado.

Abstract

The glass high voltage insulators are, today, the main equipment used in transmission lines. Although a trend towards replacing them with polymers, are still predominantly found mainly in transmission networks for high voltage 69 kV, 230 kV and 500 kV. Located over an extensive geographical area, the methods and inspection procedures become difficult to perform. The routines of preventive and predictive inspections performed today are equipped with enough subjectivity and have many factors that restrict its implementation. Depend solely on the visual assessment of the technical and specific climatic conditions: night, minimum source of light, especially the moon, high relative humidity and the use of binoculars. Other inspection techniques were studied, but have not found reports of their use in energy companies, for the mentioned tensions. This research presents a method for carrying out inspection on high voltage insulators. The differential method is the use of set in signal processing techniques with artificial intelligence, so that the electromagnetic spectrum radiated by the equipment when in operation, should be interpreted and classified according to the level of pollution of the insulators. Thus, an objective criterion is inserted in the process, giving the inspector a more efficient tool that does not require only a subjective assessment. Artificial neural networks performed the classification of signals after processing these through the wavelet transform. Measurements in the laboratory and field tests were conducted to serve as a database for training method. The best results show up hit 96.5% in all measurements.