The application of the wavelet packet to the study of vibration of transformer under DC magnetic bias

Abstract

In this paper, via analyzing the characteristic of WPT, a method of obtaining the energy eigenvector of transformer vibration based on wavelet packet transform is brought forward; the eigenvector can be used for determining the relationship between vibration levels and influence degree of DC magnetic bias to power transformers, providing criterion for analyzing the generating reason of transformer vibration and noise. At the lab, using the transformer DC magnetic bias monitoring device, vibration signals of transformer with no-load were tested under DC bias. The vibration signals in different magnetic bias level were analyzed using WPT and the energy eigenvectors were acquired. According to the experiment, we can see this method can evaluate the magnetic bias extent of transformer effectively.