Semi-analytical and Data-Driven Models of Electric Field Strength Considering the Actual Structure of Medium Frequency Transformers

The insulation of medium frequency transformers has become an important bottleneck limiting its power density and reliability. In order to ensure the reliability of insulation, it is necessary to accurately model the local maximum electric field strength. However, winding shape and geometric parameters will affect the local maximum electric field strength of the transformer, so it is difficult to perform accurate physical analytical modeling based on Maxwell's equations. In this paper, the influence of complex factors such as winding geometry parameters on the calculation of local maximum field strength is taken into account by two types models, semi-analytical modeling and data driven modeling. The proposed semi-analytical model simplifies the problem by finite element analysis and calculates the maximum electric field strength by iterative mirroring method, which is more suitable for the case of circular windings and densely wound in primary windings. While the data driven model is suitable for any transformer geometry structure. Compared with finite element analysis, the semi-analytical model realized high calculation accuracy with 2.46% average error and less than 16% error, and the data-driven model realized high calculation accuracy with less than 5% error. In addition, the average calculation time of two models is improved by two orders of magnitude compared with FEA.