An Alternative Pre-Processing Technique Applied to Power Quality Disturbances Identification

Abstract

This research presents the development of an alternative pre-processing technique of signals based on the fractal dimension calculation, entropy and signal energy that will be applied to disturbances classification occurring in an electrical power system (EPS). With respect to the power quality disturbances, the voltage sags, voltage swells, oscillatory transients and interruptions were considered for this application. In order to test and validate the proposed technique, a representative database has been obtained through computer simulations of a real EPS using the ATP software. Through windows data and the pre-processing technique proposed, the data were directed to Artificial Neural Network (ANN) architecture to classify the power quality events. The results shown that, combining both the data pre-processing techniques and ANNs, a satisfactory performance of all the proposed methodology can be obtained.

Analysis of Wavelet Transform applied to the segmentation of disturbance signals with different sampling rates

“…However, the improvement of power quality is, usually, based on techniques applied to detection [2][3], classification [2][3][4][5][6] and location of disturbances [3]. In this sense, continuous recording of disturbance signals are required, which leads to a huge volume of data to be primarily storage.…”

“…The use WT was justified by their ability to filter noise and for precisely detect abrupt changes and discontinuities in electrical signals [4], [7]. it is important to mention that the main reason for its success lies on the fact that the Wavelet done a time-frequency decomposition.…”