PREDICTION OF THE DISTRIBUTION TRANSFORMER TEMPERATURE FOR INSIDE AND UNDERGROUND TRANSFORMER CENTRES

Abstract

ABSTRACT: Distribution transformers in the transformation centres, which produce thermal losses, are closely linked with their correct proper operation, maintenance and life of these transformers. In a practical way it would be useful to estimate the temperature of the transformer from easily measurable data such as its outside temperature, the ground temperature (in underground transformers) and the power supplied by the transformer. The aim of this research is to model the behaviour of a transformer (either inside or underground transformer) in order to predict the temperature of the transformer, to optimize operating system installation and to extend the life of the machine The methods proposed in this study for such a model are artificial neural networks (RNA) and least squares polynomial regressions (RP). There were used three transformation centres for data collection. Keywords: distribution transformer temperature, loss of life, integral transformer substation, artificial neural network, polynomial regression.