Parameters identification of sectional winding high frequency transformer model using neural network

In this paper, the parameters of the power transformer model are identified to simulate its behavior under high frequency transients. The mathematical description of the sectionalized model of the transformer using two-port network topology is given. The adoption of the neural network to estimate the parameters of the sectionalized model is presented. The proposed estimation method overcomes the assumption of identical sections. The results show the ability of the proposed method to estimate the parameters with less sensitivity to the initial guess of the parameters.