Efficient Numerical Modeling of Oil-Immersed Transformers: Simplified Approaches to Conjugate Heat Transfer Simulation

Abstract

The development of digital twins for power transformers has become increasingly important to predict possible operating modes and reduce the likelihood of faults. The accuracy of these predictions relies heavily on the numerical models used, which must be both simple and computationally efficient. This work focuses on creating a simplified numerical model for a template oil-immersed power transformer (100 MVA, 230/69 KV). The study investigates how the number of elements and the strategies used to set up the mesh in the domain of interest influence the results, aiming to identify the key parameters that affect the outcomes. Furthermore, a significant effect of resolving thermal boundary layers on the accurate identification of hot spots is demonstrated. Two approaches to resolving thermal boundary layers are explored in this work. This study presents a comprehensive analysis of three numerical models for conjugate heat transfer simulations, each with distinct features and computational domain compositions. The results show that the addition of extra calculation domains leads to the emergence of new vortex structures, affecting the velocity profile at the channel inlet and altering the location of hot spots. This study provides valuable insights into the configuration and composition of calculated domains in numerical models of oil-immersed power transformers, essential for the accurate prediction of hot spot temperatures and ensuring reliable operation.