Diagnóstico de falhas em transformadores de potência utilizando sensores de gás semicondutores

Abstract

This dissertation introduces the study of the use of tin dioxide gas sensor array applied to fault detection in power transformers, through dissolved gas-in-oil. The aim of this research is to analyze the data obtained from gas sensors array, using pattern recognition techniques, and thus to determine if the use of gas sensor array allows the identification of faults. In order to facilitate the understanding of this research, the document was organized covering the following topics: power transformers, mineral insulating oil, analysis and interpretation of dissolved gás-in-oil, tin dioxide sensors and data acquisition instrument, techniques of pattern recognition and methodology for generating faults. In the end of this dissertation, the data from the sensor array are analyzed, in order to achieve the objective of verifying its application in the identification of faults.