Characteristic Analysis and Optimization Strategy of Transformer Vibration Signal

Transformer vibration analysis method often ignores the difference and connection between time domain, frequency and other characteristics. There is no comprehensive and effective extraction of vibration signal characteristics, and some important features may be lost, which may lead to poor evaluation results. In this paper, the characteristics of measured signals are analyzed by general statistical characteristics and distribution characteristics and wavelet packet transform. By comparing the differences of each characteristic parameter, the characteristic parameters with large discrimination and favorable for transformer condition assessment are selected, which lays the foundation for evaluating the mechanical stability of transformer. Finally, the experimental platform is built to verify the rationality and effectiveness of the method and feature parameter selection.