Transformer Winding Analysis Based On Finite Element Method And Parameters Identification

Transformer is an important device in power system,and according to the actual operation of power system,transformer winding deformation resulting from internal turn-to-turns faults has risen to the first place on the grid accident rate,which does serious damage to power grid security.After transformer winding deformation,its physical size and geometry shape change,and winding leakage magnetic field,as well as leakage inductance.It affects the safe operation of the transformer.So,it is of great significance to analysis transformer leakage magnetic properties and deformation as well as corresponding parameters identification.This article use traditional method to analysis the transformer leakage magnetic field distribution.After reasonable simplified,PDE Toolbox in MATLAB is used to establish transformer model and select appropriate section to obtain a two-dimensional actual transformer leakage magnetic field distribution.At the same time,the axial and radial leakage magnetic field of winding are extracted and analyzed separately.In addition,leakage inductance and magnetic energy are calculated by finite element method.Compared with traditional method,the result verify the correctness of the finite element method.This article analysis the reason of transformer winding deformation.It propose to analysis the change of magnetic field distribution and leakage inductance values after different amounts and different type of winding deformation.The normal axial and radial deformation is simulated.The result verify that magnetic field distribution and leakage inductance is affected by winding geometry and spatial location directly and the properties of different deformation is different.Due to the relationship between the winding deformation properties and parameters,this article analysis the identification number of different operational situation and present a recursive least squares algorithm with dynamic factor.It propose to use the variance to judge the volatility of recognition results and the effect of load factor and power factor of the recognition results is verified.Simulation results in MATLAB prove that identification result of recursive least squares algorithm is stable and fast.It can identify resistance and leakage inductance of the transformer intuitively and accurately and understand the transformer operation.