A multi-scale thermal-fluid coupling model for ONAN transformer considering entire circulating oil systems

Abstract

The cooling ducts of radiator play an essential role in transformer heat dissipation, while it is always ignored in most researches due to the computation costs, which might result in misleading thermal insulation margin. This paper presents a novel method, the multi-scale thermal-fluid coupling, based on solving models in different dimensions combinedly, which can accurately analyze transformer temperature with the entire circulating oil system. Firstly, finite volume method (FVM) is adopted to investigate the thermal performance of 3D winding model while the radiator cooling duct model is simplified to 1D based on nodal data. Then, a serial coupling method is utilized to map temperature and velocity data between two models, that’s to say, 3D winding model and 1D cooling duct model are connected and solved simultaneously. After the theoretical analysis, a 35 kV oil-natural-air-natural (ONAN) transformer is utilized to carry out temperature-rise test. In measurement, dynamic temperature information of winding and radiator is recorded by fiber optic sensors (FOSs) and thermocouple probes respectively. It shows that the multi-scale model on the basis of entire circulating oil systems is in accordance with the experimental results. Finally, a traditional model without the radiator cooling ducts is constructed as a comparison to demonstrate the effect of radiator cooling ducts. The results indicate that the radiator model makes a great contribution to temperature distribution, and it can change the peak value and location of hot spot. Therefore, the multi-scale model, which can combinedly solve models in different dimensions, is a more accurate and efficient method in thermal-fluid coupling, and it can offer guidance to the transformer thermal insulation design.