A fast calculation method for magnetic field of 110 kV oil-immersed transformers based on the combination of order reduction and response surface methodology

Abstract

To solve the problem of slow simulation calculation of magnetic field in oil-immersed transformers, this paper proposes a fast calculation method based on the proper orthogonal decomposition (POD) and Kriging response surface method. First, based on the POD reduction method, the discrete equation of the magnetic field is reduced to obtain the modal coefficient matrix. Then, taking the transformer load rate as an important parameter, the sample space is determined, and the corresponding snapshot matrix of the sample space is obtained through COMSOL software. Based on the Kriging function, a response function for the load rate and modal coefficients is constructed, and combined with the reduced mode, the winding magnetic field is quickly reconstructed. Finally, to verify the effectiveness of the method, four additional load conditions were selected to analyze the deviation between the reduced-order model and the full-order model from the perspective of computational accuracy and efficiency. The results show that, in terms of computational accuracy, the maximum relative error between the reduced-order model and the simulation model in the verification points is only 1.045%, while the rest of the points are below 1%. In terms of computational efficiency, the method proposed in this article improves the computational efficiency to 4743 times that of the full-order model. It only takes 0.9 s to calculate the results corresponding to a certain operating condition, which can achieve fast calculation.