The Research On Establishing Method Of The Nonlinear Broadband Model For 10kV Power Transformer

As the key component of grid, transformer is one of the most important and difficult point in the grid researching. And domestic and foreign scholars don’t terminate to research about the power transformer modeling. So for, few models can simulate a variety of conditions in a power transformer’s response. As some of the transformer parameters are both nonlinear and frequency dependent. The research of nonlinear broadband model of power transformer is put forward based on above problems. Therefore, it is necessary to establish a new broadband nonlinear transformer model.The paper studies the establishing method of nonlinear broadband transformer model, and based on a 10 k V experiments transformer set up a circuit simulation model. At first, it summarizes the methods of transformers modeling. Besides, the electromagnetic phenomena of the power transformer are analyzed in wide frequency.Secondly, the circuit which consists of a linear module and a nonlinear module is derived by the analysis of the basic structure of the transformer. Based on “Swept frequency method” and “Scattering parameter measurements”, the linear admittance parameters are measured. And then, the linear module is established by modified vector fitting and circuit synthesis. The admittance of iron core is derived by mathematical calculations of the silicon steel. The formula of frequency dependent is matched by modified vector fitting in this paper, and though compared with existing methods the modified vector fitting method is more accurate and converge faster. After circuit synthesis, the nonlinear of transformer is added in the circuit model as a nonlinear inductance.At last, a 10 k V experiment transformer is modeled, and the circuit model is simulated in software Simulink. To verify the model, the experiments of power frequency steady state, power frequency overvoltage and lightning overvoltage are conducted. The comparisons between the simulation and measurement results verify the feasibility and validity of the model.