Answers are grounded in this paper and cite their sources.
This computational strategy improves hot-spot temperature estimation and reduces errors in oil-immersed transformers.
To enhance the accuracy and efficiency of transient temperature rise and hot-spot temperature calculations for oil-immersed transformer windings, this study proposes an extended dynamic mode decomposition computational strategy. This method incorporates dictionary functions for dimensionality mapping, effectively addressing the significant increase in computational errors associated with long-term strong nonlinear predictions in Dynamic Mode Decomposition (DMD). By selecting the dominant modes, the computational efficiency is improved, ultimately leading to high-precision and rapid prediction of the full-field temperature based on historical simulation data. This study discusses the performance of 16 different dictionary functions to identify the optimal combination. A two-dimensional simulation model and a temperature rise experimental platform were established based on the fundamental structure of a product-grade 110 kV oil-immersed transformer winding to validate the proposed method. The results indicate that the calculation results of this method are approximate to those of full-order computations, with a maximum hot-spot error ∼1/5 of that in DMD, and the computation time reduced to only 1/29 of that required by Fluent. For the temperature rise experiment, the average hot-spot error is within 0.60 K, and the average computation time of per time step is only 0.14 s. The findings demonstrate that the proposed method ensures the accuracy and efficiency of transient temperature rise calculations, offering significant application value in the rapid simulation calculation and digital transformation of transformers.
No takes yet. Share an insight, caveat, or question.
Liu et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: