Rede Neural ARTMAP-FUZZY e Transformada Wavelet para Detecção e Classificação de Distúrbios de Tensão em Sistemas de Energia Elétrica

− Many efforts have been dispended to solve problems related to the Electrical Energy Quality, principally in automation of process and developing monitoring equipments that provide improvements in behavior and reliability of the electrical system. This paper presents an automatic identifier/classifier system for disturbances called Wavelet-ARTMAP-Fuzzy neural network. The basic structure of this neural network is composed of three modules: the anomaly (disturbance) detection module; the characteristics extraction module, where the wave forms are analyzed by calculating the Discrete Wavelet Transform, Multiresolution Analysis, and Entropy Norm; and the classification disturbance module which contains a Fuzzy ARTMAP neural network that shows what kind of anomaly of the signal. This study considers seven types of electrical signals, generated from the mathematical models, performing 2800 wave forms. Thus, the performance of this network in detecting and classifying correctly the several electrical disturbances was 100%, moreover the robust form and velocity in obtaining the results, allowing using in real time.