Procedimiento para el análisis de calidad de potencia basado en Wavelets y Redes Neurales

This work describes a complete procedure for the analysis of power quality, starting from established formsthrough electrical disturbances standards. For it, disturbed signal is decomposed in different levels ofresolution through the Transformed Wavelet (WT), and then, significant indicators are obtained by anadequate statistical processing, which permit the classification of the different disturbances. The aboveclassification is made (from the indicators used as inputs) using a neural network, which is trainedaccording to a Back-propagation technique, achieving satisfactory results under a serie of simulatedadverse conditions.