Numerical investigation of oil flow and temperature distributions for ON transformer windings

In this paper, numerical investigation of oil flow distribution and temperature distribution is performed for a disc-type transformer winding in an oil natural (ON) cooling mode. First, dimensional analysis is carried out based on the governing differential equations with the Boussinesq approximation, and Re, Ri and Pr are found to be the governing dimensionless parameters in determining flow and temperature distribution. Then a CFD case study is performed on a winding model as the benchmark for this study, in which reverse flow and hot-plumes are observed. After the case study, CFD parametric sweeps of Re and Ri are executed. It is found that the minimum value of the hot-spot factor, which characterizes the thermal performance of the transformer winding, is achieved in a relatively small and fixed range of Ri (from 0.4 to 0.6) regardless of the values of Re and Pr in their practical ranges. Consequently, the relationship of an invariable minimum hot-spot factor with a small fixed range of Ri makes the optimization of the transformer operational regime possible. Finally, a new CFD case study is performed to confirm the shifting of an uncontrolled operational regime to a quasi-optimal one by changing Ri from 1.46 in the benchmark case to 0.6 in the new case.