Services on Demand
Journal
Article
Indicators
- Cited by SciELO
- Access statistics
Related links
- Cited by Google
- Similars in SciELO
- Similars in Google
Share
DYNA
Print version ISSN 0012-7353On-line version ISSN 2346-2183
Abstract
CALDERON, JHON ALBEIRO; MORENO CADAVID, JULIAN and OVALLE, DEMETRIO ARTURO. NEURAL NETWORK FOR FAULT CLASSIFICATION IN TRANSMISSION LINES BASED ON OSCILOPERTUBOGRAPHY RECORDS. Dyna rev.fac.nac.minas [online]. 2008, vol.75, n.156, pp.99-107. ISSN 0012-7353.
The electric fault diagnostics in high voltage transmission lines is a complex task not only because of the amount of information which can come from different sources as SOE, SCADA and recorders, but also because of the variability of such faults. Such complexity impacts the opportunity and accuracy of diagnostic, and these issues are particularly important for actual time analisis where adecuate actions should be taken in order to restablish the electric power system. In this paper a neural network with bayessian regularization learning and early finalization is proposed for fault classification ussing osciloperturbography records and its efectivity is shown for a wide variety of training and validation cases which are obtenied with an ATP model where the required electric faults were simulated.
Keywords : Power System; Fault Diagnostics; ATP; Neural Networks; Bayessian Regularization.