Decomposed fixed-point harmonic-balance finite element method and its application in DC bias phenomenon

This paper proposes a decomposed fixed-point harmonic-balanced finite element method (DFPHBFEM) to analyze the nonlinear time-periodic magnetic field in the power transformer. Based on the fixed-point reluctivity, the constitutive relation H-B-M is introduced to establish equation, then locally convergent algorithm (LCA) is used to determine the fixed-point reluctivity in each finite element. The block Gauss-Seidel algorithm and the relaxation iterative method are used to solve the equation. Via the above methods, the nonlinear magnetic field can be solved accurately and efficiently, moreover DFPHBFEM greatly reduces the demand for computer memory, and makes it applicable to the large-scale computation of actual power transformers. DC-biasing experiments are carried out on the laminated core model (LCM), and the proposed algorithm is applied to calculate the nonlinear magnetic field and the exciting current. Comparing the calculated results with the experimental data shows consistency, which proves the effectiveness of DFPHBFEM.