Research On Condition Assessment Of Power Transformer Based On Vibration Signal Analysis

In recent years,China has built a large number of ultra-high voltage substation to improve the voltage level of the power grid.As the key equipment of the power grid,it is more and more important to ensure the safe and stable operation of the transformer.In this paper,the relationship between the magnitude of preload and the acceleration of vibration signal is analyzed theoretically and the distribution law of the natural frequency at different axial positions of windings is studied.Besides,based on the ensemble empirical mode decomposition and the negative entropy criterion,a new method is proposed for the blind separation of single chann el transformer vibration signal,and the health status of transformer is evaluated based on the vibration signal data.The details are as follows:Based on the spring mass model of the axial vibration of winding,the relationship between the magnitude of the preload and the amplitude of the vibration acceleration is analyzed.According to the influence of the winding’s own gravity and the steady state component of the ampere force,the force distribution of the windings is analyzed and the distribution law of axial natural frequency is obtained.Then the correctness of the theoretical analysis is verified by analyzing the vibration signal data of the actual transformer.A blind source separation method of transformer vibration signals based on ensemble empirical mode decomposition and negentropy criterion is proposed in this paper.Compared with the traditional method of blind source separation,the method for which the underdetermined blind source separation is transformed into the adaptive blind source separation realizes the separation of every vibration source signal from the single channel mixed vibration signal by raising the dimension.The validity of the proposed method is verified through the analysis of the simulated and measured signals and the comparison of the results of Fast ICA separation.A method of state estimation based on the spectrum distribution characteristics of multiple points is proposed,and the spectrum distribution characteristics are described by the frequency complexity and the percentage of energ y distribution in different frequency bands to analyze transformer vibration signal spectrum distribution characteristics.By calculating the correlation of the energy distribution percentage of each frequency band,the transformer operation state is evalu ated.