Modeling Power Transformer Field Drying Processes

Abstract

Power transformer field drying is becoming a habitual practice because water is damaging for transformer insulation, and its presence increases the probability of unexpected failures. Different drying methods are currently being used by electrical companies but sometimes without a profound knowledge of the drying processes involved, and consequently drying is not as effective as would be desirable.

Physical models have been developed by the authors in order to study power transformer drying processes. The use of the models will help to plan more effective drying processes, tailoring the drying times and drying conditions for each particular case. The models have been tested in a test transformer fitted with sensors. In order to monitor the process, insulation samples were extracted from the transformer before and after the process.

In this article, several transformer drying procedures are described. Theoretical models and their experimental validation processes are reported. Finally, some general recommendations about transformer drying in the field are provided.

Keywords:

• Drying time
• Field drying
• Hot oil drying
• Low-frequency heating
• Power transformer
• Transformer insulation
• Vacuum drying

ACKNOWLEDGMENT

This work was financed by the Spanish Ministry of Science and Technology under the project DPI2009-7093.

Notes

¹The Schmidt number and the Lewis number represent, respectively, the ratio between the momentum diffusivity and the mass diffusivity and the ratio between the thermal diffusivity and the mass diffusivity.

²The probes are able to determine the temperature with a resolution of 0.1°C.

³Temperature measurement started some hours later than the first period of drying; therefore, only 14 h of temperature measurements are registered during the vacuum drying period (Fig. 6).

⁴According to UNE-EN 60814-1999,^{[ 25 ]} two measurements of moisture content were carried out over every sample. The moisture contents displayed in Table 1 were obtained as the average value of these measurements rounding off to 0.01% by weight.

⁵After these experiments an automatic temperature regulation system was implemented.