AGRADECIMENTOSAgradeço primeiramente a Deus pela minha vida e pelo trabalho realizado.Ao Prof. Dr. Mário Oleskovicz, meu orientador, pelos ensinamentos, orientações, incentivo, paciência que muito contribuiu para meu crescimento intelectual e científico. all the n possible disconformities in the waveform that can exist or occur in the data file or registered data in analyses. As the differential of this research, from the location in time of the alteration/discontinuities in the waveforms, it was possible to obtain data windows with flexible sizes. So, several events on the signal were evaluated, as well as their specific timeduration. For this purpose, the Wavelet Transform (WT) was used to reach the detection and localization in time of the waveform alterations. For the event classification, the TW, the Fourier Transform (FT) and the Root Mean Square (RMS) value were used. The data window flexibility allowed an appropriate choice of which tool could be better used in the classification task. As implemented, each one of these tools presented an answer, and the final answer was obtained by using a logic decision module. To validate the study, some situations of disturbances were characterized using a real distribution system, implemented and simulated applying the ATP (Alternative Transients Program) software. The results were excellent in such a way for detection and localization in time, as well as for the automatic classification and estimation of the magnitude and the event duration.