Análisis de fallas en transformadores de potencia de centrales de generación eléctrica

Abstract

Power transformers have particular design conditions that are a function of many factors, the following are the most relevant ones: factors of environmental conditions at the installation site, load regime, derating factors, derating factor at different heights , degree of salinity, etc. Consequently, it is necessary to apply different maintenance techniques depending on the exposed conditions. In this context, the purpose of this development project is to apply a non-standard technique based on Artificial Neural Networks (ANN) for the interpretation of the different levels of gases contained in the dielectric oil (only of mineral origin), contributing to the decisions made in the maintenance programs that are implemented in the power transformer of the power plants. generation.