A fast computational method for internal temperature field in Oil-Immersed power transformers
Gang Liu,Wanjun Hu,Shiyuan Hao,Chenglong Gao,Yunpeng Liu,Weige Wu,Lin Li
DOI: https://doi.org/10.1016/j.applthermaleng.2023.121558
IF: 6.4
2023-09-12
Applied Thermal Engineering
Abstract:In order to improve the computational efficiency of the steady-state temperature rise of oil-immersed power transformer windings, this paper proposes a temperature field reduced order model(ROM) based on proper orthogonal decomposition (POD) and response surface method(RSM). The method takes the operating conditions of the transformer as input and uses the POD method as a carrier, and establishes a fast calculation model of the steady-state temperature field through RSM, thus achieving fast calculation from transformer operating data to the entire temperature field. The paper first analyzes the reduction characteristics of the POD method, and then introduces the RSM to establish the correlation between the POD modal coefficients and the transformer conditions. Compared with traditional surrogate models, this method can obtain high-precision surrogate relationships with a small number of sample points, and the time cost of establishing is relatively low. The method aims to quickly obtain the POD modal coefficients through the winding conditions, thereby skipping the complex nonlinear calculation and efficiently reconstructing the winding temperature field with the reduced modal. The related examples show that the method has good computational accuracy and efficiency. In a 2D forced oil circulation cooling model, out of 100 test conditions, the error is not more than 2.0 K compared with the full-order calculation, and the total calculation time is only 3.13 s. Finally, based on a 35 kV natural oil circulation cooling transformer, a temperature rise test platform is constructed to verify the effectiveness of the method. The experimental results show that the calculated error of the ROM is not more than 5 K compared with the experimental results, and the single-step calculation time is only 0.73 s, which has greatly improved the computational efficiency compared with full-order calculation of the same scale. The proposed method in this study achieves a second-level calculation efficiency for the transformer winding temperature, providing a feasible solution for digital online monitoring of transformer temperatures.
energy & fuels,engineering, mechanical,thermodynamics,mechanics