Robust features selection scheme for fault diagnosis in an electric power distribution system

Author

Butler, Karen L. ; Momoh, James A.

Author_Institution

Dept. of Electr. Eng., Howard Univ., Washington, DC, USA

fYear

1993

fDate

14-17 Sep 1993

Firstpage

209

Abstract

In this paper, the authors present a statistically derived features set for use as input to a neural network based arcing line fault detector for power distribution systems. In addition, the authors test the performance of the back-propagation artificial neural network uses values of the features set computed from laboratory experimental data. The results show great promise toward the development of an efficient artificial neural network based arcing line fault detector

Keywords

arcs (electric); backpropagation; distribution networks; fault location; neural nets; pattern recognition; arcing line fault detector; back-propagation; feature extraction; kurtosis; neural network input; pattern classifier; power distribution systems; reflection coefficients; robust features selection scheme; skewness; statistically derived features set; Artificial neural networks; Electrical fault detection; Fault detection; Fault diagnosis; Neural networks; Phase detection; Phase frequency detector; Power distribution; Robustness; Solids;

fLanguage

English

Publisher

ieee

Conference_Titel

Electrical and Computer Engineering, 1993. Canadian Conference on

Conference_Location

Vancouver, BC

Print_ISBN

0-7803-2416-1

Type

conf

DOI

10.1109/CCECE.1993.332293

Filename

332293