Soundprint Feature Analysis of Main Transformers in a 500kV Substation

This paper presents the transformer soundrint characteristic quantity analysis method, and then selects a 500kV transformer as the research object. Determine the 500kV transformer from day 19 to day 20. In the case of DC magnetic bias, the system measures the same characteristic trend from A, V, and C three phases, and makes an accurate judgment. Therefore, comprehensive multi-point data collection and analysis can make a better judgment of the DC partial magnetism of the power grid. If the soundprint monitoring linkage is realized in different substations of the power grid, it should be able to be very accurate. The voice print characteristic quantity threshold should be set according to the characteristics of the transformer. At present, the threshold value of the transformer adopts the default value, which is not adjusted according to the actual situation of each transformer, resulting in the deviation of the system judgment state. Considering this analysis, the feature quantity can be adjusted in the following ways, such as the harmonic ratio should be raised appropriately, and the complexity threshold should be lowered. This requires constant tracking, appropriate adjustment of the number of features, to achieve accurate fault alarm, and less false alarm..