Thermal management of a distribution transformer: An optimization study of the cooling system using CFD and response surface methodology

Abstract

In this paper, a numerical scheme has been developed to examine the effective parameters on thermal management of distribution transformers and subsequently to optimize their cooling systems. In this regard, the response surface methodology (RSM) was used as the optimization method to minimize the hotspot temperature in the transformer as the response factor. A comprehensive three-dimensional computational scheme was employed considering the detailed geometrical specifications of an actual 200kVA distribution transformer to obtain the temperature field and the hotspot temperature. The accuracy of the numerical model was established via comparing the numerical results with the measured temperatures of a running transformer. The thermal variations of the thermo-physical properties of the transformer oil are determined experimentally and incorporated in the numerical modeling. A comprehensive parametric study among seven evidently effective parameters has been performed to identify the most effective parameters on the thermal performance of the transformer. It was found that fin height, length, and spacing are the more influential parameters among the examined parameters, which are also considered as the input variables in the optimization procedure. According to the RSM, the effects of the variations of these input variables in pre-specified ranges on the response, which is the hotspot temperature, are examined through the suggested runs by RSM. The results indicated that the hotspot temperature is more influenced by the fin height as compared to the fin length and spacing. Furthermore, the hotspot temperature decreases with the increase in fin height and length, while decreases as fin spacing increases. In addition, a correlation for the variations of the hotspot temperature as a function of the fin height (H), length (L), and spacing (S) is suggested using RSM. The significant finding is that the proposed optimum transformer configuration (H = 0.9 m, L = 0.08 m and S = 0.036 m) leads to the hotspot temperature reduction of about 16 °C as compared to the actual transformer geometry, which greatly affects the transformer life expectancy and safer performance.