Prediction of core losses on a three-phase transformer using neural networks

This work is a study of intelligent tools for forecasting losses in the transformer core, comparing with the traditionally accepted model. The uses of Artificial Neural Networks (ANN) search a best estimator of losses in the transformer core. Two models of ANN were proposed: Multi-Layer Percepton (MLP) and Neo-Fuzzy Neuron (NFN). After an analysis of the model they were evaluated and it was concluded that the MLP model for this application had the best performance.