Analysis of long-term monitoring of autotransformer neutral currents

• Main Authors: Selivanov V. N., Barannik M. B., Bilin V. A., Efimov B. V., Kolobov V. V., Sakharov Ya. A.
• Format: Article
• Language: Russian
• Published: Murmansk State Technical University 2018-12-01
• Series: Vestnik MGTU
• Subjects: geomagnetic storm; geomagnetically induced current; overhead transmission line; digital signal processing; wavelet transform
• Online Access: http://vestnik.mstu.edu.ru/show-eng.shtml?art=1984

The Kola Science Center of the Russian Academy of Sciences together with the Polar Geophysical Institute is carrying out research concerning the effect of geomagnetic storms on the state of electrical networks and transformer substations on the Kola Peninsula and in Karelia. A regional system for monitoring currents in transformer neutral was created which collected a huge array of information on the influence of geomagnetic disturbances on a 800 km electric network. Overhead transmission lines are giant antennas that capture electromagnetic energy of any nature from the surrounding space. An analysis of the data accumulated over many years shows that in the current of a grounded neutral there are reliable responses to a wide range of electromagnetic phenomena on the Earth and in the near-Earth space: pulsed and quasisinusoidal pulsations of the geomagnetic field, extremely-low frequency components of artificial signals, lightning discharges and internal transients in the power system. To identify phenomena, a visual method is used to analyze the time and spectral characteristics of signals, which takes a long time and is rather subjective. There is always the possibility that some important, previously unidentified phenomena will go unnoticed. The paper describes the characteristic features of signals detected in neutral currents of power transformers. Algorithms of digital signal processing that can be used to create programs for automatic analysis of large data sets have been considered. To the greatest extent, numerical methods based on wavelet expansion and Fourier transform are suitable for studying nonsteady signals of a complex structure.