Abstract: In the paper a new neural network based out-of-step protection scheme is presented. Numerous ANNs have been trained and tested with input patterns calculated from voltage and current signals generated with use of EMTP/ATP programme. Optimal selection of decision signals for ANN feeding has been done with help of proposed statistical PDF distance indices. The OS detection/prediction efficiency was assessed for various ANN sizes and data window lengths. The impact of measurement rate on the classification results has been analysed. Wide robustness checking of developed solutions has been done against different fault types and synchronous machine ratings.

Keyword(s): neural network ; ANN ; statistical distance indices ; probability density function ; EMTP ; power system
Note: Proceedings of the 12th International Conference on Power System Protection. PSP 2000, Bled, Slovenia, September 27th-29th, 2000 / [Ed. by Tadeja Babnik, Ferdinand Gubina]. Ljubljana : Faculty of Electrical Engineering, 2000. pp. 51-56.


Fulltext : http://zet10.ipee.pwr.wroc.pl/record/194/files/

Cited by: try citation search for ENY-ARTICLE-2008-152; I08/2000/I-133

Ähnliche Datensätze

 

Benutzer, die diese Seite angeschaut haben, haben auch angeschaut:

(943)  Impact of Distributed Generation on Electrical Power Network - Khan, U N - ENY-ARTICLE-2008-142 (942)  Proceedings of the Int. Conference "7th EEEIC International Workshop on Environment and Electrical Engineering" - Schwarz, H et al - ENY-PREPRINT-2008-008 (934)  Adaptacyjne sieci neuronowe w zastosowaniu do identyfikacji stanu układów elektrycznych - Kostyla, P - ENY-THESIS-2008-005 (921)  Zadania 1 Podstawy Elektrotechniki Automatyka i Robotyka - Hejke, I et al - ENY-TEACHING-2008-020 (904)  Measurement of IEC Groups and Subgroups Using Advanced Spectrum Estimation Methods - Bracale, A et al - ENY-ARTICLE-2008-117

 

Dieses Dokument bewerten:

Als Erster ein Review für dieses Dokument verfassen