An application of artificial neural networks (ANNs) to distance protection is presented in tius paper. A neural network was truined by data from simulation of a simple system under load and fault conditions, tested by data with different system conditions, and fmally run for faults along the whole line. The research was concentrated on creating more selective arcing fault detection, especially for radial distribution lines where arc resistance can be a significant part of the zero sequence impedance. A nonlinear arcing resistance model was used to provide data and a new operating characteristic was devised. The prospective ANN distance relay showed very good performance in detecting a single-line-to-ground fault with nonlinear arcing resistance along the whole transmission line.
[1]
A. P. Strom.
Long 60-Cycle Arcs in Air
,
1946,
Transactions of the American Institute of Electrical Engineers.
[2]
G. Swift,et al.
Detection of high impedance arcing faults using a multi-layer perceptron
,
1992
.
[3]
J. R. Dunki-Jacobs,et al.
The Effects of Arcing Ground Faults on Low-Voltage System Design
,
1972
.
[4]
Mark W. White,et al.
A neural network approach to the detection of incipient faults on power distribution feeders
,
1990
.
[5]
L. van der Sluis,et al.
Adaptive distance protection of a double-circuit line using artificial neural nets
,
1994
.