Detección y clasificación de fallas en líneas de transmisión eléctrica utilizando el método de la transformada S

Abstract

The present research focuses on the detection and classification of faults in the transmission system. For this purpose, an association criterion based on an artificial intelligence method and on the calculation of short-circuit studies of various types is used. In this sense, a database containing the magnitudes of voltages and currents generated by the shortcircuit studies is built. Subsequently, the Stransform is calculated for the analysis of the waves in the frequency domain and, consequently, the values obtained are associated with the corresponding binary vector. The binary array consists of 4 elements, one for each phase. Finally, the ANN is trained with the obtained data and proceeds to the evaluation under different fault conditions. Therefore, the 9-node test system is chosen to test the proposed methodology. The results show that the ANN together with the S-transform describes 97% efficiency for fault classification.