A study of characteristic extraction for transformer oscillation wave based on DWT analysis

Abstract

Abstract In order to improve the reliability of transformer winding state evaluation method, the analysis of oscillation wave based on high voltage direct current (HVDC) power supply is proposed. Firstly, different noise reduction methods are adopted dealing with the measured oscillation wave data, and a suitable noise reduction method for field experiments is obtained. Then the discrete wavelet transform (DWT) is applied on denoising data. Finally, the characteristics that can be used to evaluate the state of windings are obtained by analysing the wavelet detailed signals. The results show that the waveform characteristics of windings obtained by discrete wavelet transform have good repeatability and can reflect the differences between different windings.