An approach based on neural networks for identification of fault sections in radial distribution systems

The main objective involved with this paper consists of presenting the results obtained from the application of artificial neural networks and statistical tools in the automatic identification and classification process of faults in electric power distribution systems. The developed techniques to treat the proposed problem have used, in an integrated way, several approaches that can contribute to the successful detection process of faults, aiming that it is carried out in a reliable and safe way. The compilations of the results obtained from practical experiments accomplished in a pilot radial distribution feeder have demonstrated that the developed techniques provide accurate results, identifying and classifying efficiently the several occurrences of faults observed in the feeder.

Keywords

Artificial intelligence, Automation, Classification (of information), Computer networks, Electric fault location, Electric load distribution, Electric power systems, Electric power transmission, Electric tools, Electronic data interchange, Feeding, Automatic identification, Industrial technologies, International conferences, Neural networks