Large Power Transformer Core Vibration Analysis And Experimental Research

Large power transformer is one of the important equipments in the power system.Its safety is essential for ensuring the reliability of power network, which is also of great significance to realize the on-line monitoring and fault diagnosis of power transformer. The transformer core is the main components and also the main fault source. The core vibration, which is produced by magnetostriction and is correlative with the core's clamping pressure and insulation, is essentially nonlinear and nonstationary. Research on transformer core vibration and fault diagnose is an important part of power transformer on-line monitoring.In this paper, the power transformer core vibration is studied. The new time-frequency analysis method, Hilbert-Huang Transform (HHT) is introduced into transformer core vibration signal analysis. Combing HHT with Fourier transform, six different types of transformer core vibration signal are analyzed on aspect of time-frequency domain. Based on the adaptive HHT technique, we propose a method for analyzing the characteristics of the core vibration. By decomposing the preprocessed vibration signal into a set of intrinsic mode functions (IMFs) through the empirical mode decomposition (EMD), the instantaneous frequencies and the energy-frequency-time distribution of the signal are derived firstly. Then, the instantaneous frequencies and the energy-frequency-time distribution are used to construct a group of eigenvectors to represent the core parameters and a distance algorithm to classify different conditions of the core vibration, based on which the nonlinear and nonstationary characteristics of the core vibration are analyzed.Experimental results show that the derived energy-frequency-time distribution of the core vibration clearly expressed the non-liner time-varying vibration caused by magnetostriction and Electromagnetic force, which contains a wealth of magnetization, structure condition and fault information. The eigenvectors and distance algorithm proposed here can characterize the magnetostriction and pressure of the core effectively. This method is useful and helpful for vibration condition monitoring and diagnosis of power transformers.