Sistema automático de classificação de distúrbios elétricos múltiplos e isolados

Abstract

In this work, an automatic system was designed for classifying single and multiple disturbances in power systems. A filter bank was used for signal decomposition, which allows the decoupling of some disturbance classes, and makes it simpler the classifier design. The proposed system was designed using simulated data and tested for both simulated and experimental data. A global classi- fication efficiency above 99 % was achieved for the data set considered.