versión On-line ISSN 0718-0764
VILLADA, Fernando; VALENCIA, Jaime A y ORILLE, Angel L. Application of Artificial Neural Networks to the Differential Protection of Synchronous Generators. Inf. tecnol. [online]. 2006, vol.17, n.5, pp. 83-90. ISSN 0718-0764. http://dx.doi.org/10.4067/S0718-07642006000500013.
This study proposes a new algorithm for the differential protection of synchronous generators based on Finite Impulse Response Artificial Neural Networks. Also, a model for a synchronous machine was developed. The model allowed simulation of internal defects in the windings of the stator using the ATP-EMTP electromagnetic transient program. This was done to generate training patterns for neural networks. The results led to the conclusion that the new schemes proposed for differential protection were more effective than those known to date. Response times of about 2.5 microseconds were obtained against internal faults, with a high capacity for discrimination between internal and external faults.
Palabras llave : neural networks; differential protection; synchronous generator; synchronous machine.