Investigation of potential factors affecting the measurement of dew point temperature in oil-soaked transformers

Moisture within a transformer's insulation system has been proven to degrade its dielectric strength. When installing a transformer in situ, one method used to calculate the moisture content of the transformer insulation is to measure the dew point temperature of the internal gas volume of the transformer tank. There are two instruments commercially available that are designed for dew point temperature measurement: the Alnor Model 7000 Dewpointer and the Vaisala DRYCAPRTM Hand-Held Dewpoint Meter DM70. Although these instruments perform an identical task, the design technology behind each instrument is vastly different.;When the Alnor Dewpointer and Vaisala DM70 instruments are used to measure the dew point of the internal gas volume simultaneously from a pressurized transformer, their differences in dew point measurement have been observed to vary as much as 30 °F. There is minimal scientific research available that focuses on the process of measuring dew point of a gas inside a pressurized transformer, let alone this observed phenomenon.;The primary objective of this work was to determine what effect certain factors potentially have on dew point measurements of a transformer's internal gas volume, in hopes of understanding the root cause of this phenomenon. Three factors that were studied include (1) human error, (2) the use of calibrated and out-of-calibration instruments, and (3) the presence of oil vapor gases in the dry air sample, and their subsequent effects on the Q-value of the sampled gas. After completing this portion of testing, none of the selected variables proved to be a direct cause of the observed discrepancies between the two instruments.;The secondary objective was to validate the accuracy of each instrument as compared to its respective published range by testing against a known dew point temperature produced by a humidity generator. In a select operating range of -22 °F to -4 °F, both instruments were found to be accurate and within their specified tolerances. This temperature range is frequently encountered in oil-soaked transformers, and demonstrates that both instruments can measure accurately over a limited, yet common, range despite their different design methodologies.;It is clear that there is another unknown factor present in oil-soaked transformers that is causing the observed discrepancy between these instruments. Future work will include testing on newly manufactured or rewound transformers in order to investigate other variables that could be causing this discrepancy.