Abstract
Faults may occur in transmission lines due to various causes, including wind, salt, rainfall, snow-fall, contact with wildlife and flying objects. When these types of faults occur, current and voltage waveforms undergo various changes depending on the particular cause. Since these changes are character-istic of the particular cause of the fault, it is possible to predict the cause of a fault based on these waveform characteristics. Therefore, we conducted artificial fault testing on an actual system scale to clarify the relationships between fault causes and changes in current and voltage waveforms.
Moreover, it was confirmed that the cause of a fault could be determined at a high rate of accuracy by the multi-discriminant and the neural network analysis. As a result of examination and evaluation of these findings, it was found that it will be possible in the future to automate the location of faults and their causes that have thus far been dependent on manual fault detection operations by maintenance personnel.
