Skoči na glavni sadržaj

Izvorni znanstveni članak

Incremental and stable training algorithm for wind turbine neural modeling

Slim Abid ; Control & Energy Management Lab (CEM LAB)National School of Engineering of Sfax, B.P. 1173, 3038 Sfax,University of Sfax, Tunisia
Mohamed Chtourou ; Control & Energy Management Lab (CEM LAB)National School of Engineering of Sfax, B.P. 1173, 3038 Sfax,University of Sfax, Tunisia
Mohamed Djemel ; Control & Energy Management Lab (CEM LAB)National School of Engineering of Sfax, B.P. 1173, 3038 Sfax,University of Sfax, Tunisia


Puni tekst: engleski PDF 438 Kb

str. 165-172

preuzimanja: 490

citiraj


Sažetak

Training and topology design of artificial neuralnetworks are important issues with largeapplication. This paper deals with an improvedalgorithm for feed forward neural networks (FNN) straining. The association of an incrementalapproach and the Lyapunov stability theoryaccomplishes both good generalization and stabletraining process. The algorithm is tested on windturbine modeling. Compared to the incrementalapproach and to the Lyapunov stability basedmethod, the association of both strategies givesinteresting results.

Ključne riječi

wind turbine; neural models; incremental algorithm; adaptive learning rate

Hrčak ID:

111105

URI

https://hrcak.srce.hr/111105

Datum izdavanja:

25.11.2013.

Posjeta: 1.008 *