| dc.description.abstract | With the integration of carbon neutral renewable energy sources and the electrification of transport and heating, transformers may face unpredictable load patterns with high fluctuations where overloading could become necessary due to economic reasons or simply to ensure continuous energy supply. A possible cause of failure from short-time emergency loading scenarios with a high temperature and rapid temperature rise is the bubble formation from transformer winding insulation. Apart from the general agreement that the water content in paper is the most significant impact parameter, there is a lack of comparability from test setups and measurement methods, as experimental results on bubble formation vary widely among different research laboratories for conditions which are claimed to be similar. Identifying possible causes and additionally quantifying impacts of influential parameters will benefit the understanding of the bubble formation phenomenon.
A bubble formation test setup that represents a transformer localised hot-spot has been developed and validated. Following a thorough thermal analysis of the setup, a test method of applying a two-stage measurement procedure was introduced. This eliminates the impact of the thermocouple placement on the bubble formation process, while it also obtains the bubble formation temperature at the location of bubble formation. Furthermore, a unique technique to measure the water content in paper immediately after the initial bubble formation has been introduced, which allows to relate the water content associated with the bubble formation process with the temperature at this time and to study the moisture migration. The first instance of a sign of a bubble was defined as the time of bubble formation as the initial accumulation of bubbles between paper layers already impacts the dielectric properties in this area. The bubble formation was investigated for different insulating material combinations, including non-thermally upgraded Kraft paper and thermally upgraded paper as solid and mineral oil, gas-to-liquid oil, and synthetic ester as liquid ones. Apart from the material type, the impact of the thermal profile with two different power inputs and the number of paper layers (two, four and eight) on the bubble formation process were investigated. An extensive amount of over 250 tested samples provides high confidence in the reported findings, which cover a water content in paper range from around 1% to 6%.
The power input and the number of paper layers did not impact the bubble formation temperature according to the water content in paper at inception, however, the moisture migration was impacted from them. This finding indicates that the bubble formation depends on an absolute temperature depending on the water content in paper at inception, regardless of the water migration before the bubble formation takes place. The test with the different liquid types showed that they have an insignificant impact on the bubble formation temperature. This implies that for the alternative-liquid-immersed transformers, in the context of bubble formation, the current IEC/IEEE loading guide for operation could be applied.
The bubble formation temperature according to the water content in paper at inception was only affected from the paper type. The thermally upgraded paper in this study has a significantly lower bubble formation temperature than the non-thermally upgraded Kraft paper at low water content values, which is 155 °C versus 189 °C at around 1% and 139 °C versus 161 °C at around 2%. For high water content values, at around 6%, the difference is smaller with only 8 K. This significant difference indicates that a revision of the IEC/IEEE standard, where no paper type is mentioned, might be necessary to cater for the individual differences. Consequently, parameters have been proposed which could be applied to the formula in the standards to take the observed differences from the tested paper types into account. | en_GB |
