Diagnosis of lightning arrester Using Fuzzy Logic and Wavelet Transform

Abstract

This paper presents an evaluation of thermal images from electrical substations belonging to the eletronorte/eletrobrás power plant, located in the Amazon region, Pará, Brazil. The solution presented in this paper is based on fuzzy logic and discrete wavelet transform. Wavelet transform decomposes thermal images, obtained by thermal imaging processes, into bands of different frequencies. The solution presented allowed for the detection of hot and not hot points, along the body of a lightning arrester, through noise discrimination according to scaling. The development of a fuzzy system solution was designed to predict average temperature differences using the thermal imaging data. By means of the proposed solution it was possible to prove that some variables observed during the capture process, of imprecise domains, can influence the generated diagnoses.