A tank vibration model for online monitoring of power transformer

In this paper, a nonlinear N-L-N Hammerstein-Wiener model is used to build the vibration model of power transformer for the online monitoring of winding condition. The vibration mechanism for the identification of power transformer is presented first. Then the Hammerstein-Wiener model is introduced and ways to estimate the parameters of the model is illustrated explicitly. Validation over an in-service gird power transformer is done to verify the effectiveness of the model. The estimated results show good fitness with the measured data collected from the vibration monitoring system, which indicates its correctness of the proposed model. Moreover, compare to existing methods, the model presented in this paper gives a more comprehensive estimation including both linear and nonlinear information of the system which is beneficial for the online condition monitoring of power transformers.