OPTIMIZING PERFORMANCE OF A TRANSMISSION LINE RELAYING ALGORITHM IMPLEMENTED USING AN ADAPTIVE, SELF-ORGANIZED NEURAL NETWORK
S. Vasilic, Mladen Kezunović, D.J. Šobajić · 2002
This paper deals with neural network pro- tection algorithm aimed at classifying transmission line faults. The approach utilizes self-organized adaptive type of neural network specially developed to deal with large amount of data. Various procedures of preprocessing neural network inputs are extensively investigated. The classification performance for different values of data window length and sampling frequency is demonstrated and illustrated. Simulation results show satisfactory algo- rithm responses in each of the implemented cases.