Time-frequency analysis for power transformer fault detection using vibration method

The vibration methods are widely employed in the machinery condition monitoring and the time-frequency analysis has been applied in many fault detection applications. In this paper, the study of the vibrations using Wigner-Ville distribution is presented to assess the power transformer condition. First, the distribution contour plots of power transformer vibrations under different conditions are compared. Then, the pattern recognition based on the energy distribution similarity is presented. Finally, a health metric is proposed to represent the transformer health state. The results from more than 10 transformers are compared, and the preliminary study shows that the proposed method is effective to assess the power transformer condition.