Dielectric Breakdown Voltage Testing - Part 5

Dielectric Breakdown Voltage Testing Part 5 – ASTM D1816 for In-service Insulating Liquids - What We Do With a Failed Test

When we evaluate D1816 results that are outside the acceptable range, the first consideration is to cross reference the results to moisture, liquid power factor, and liquid screen tests to identify possible causes for the depressed D1816 values. The dissolved gas analysis is also consulted to determine whether the gas content could have depressed the D1816 to be outside the acceptable range. Frequently, we can eliminate moisture, oil aging, and most types of contamination as probable causes for poor D1816 results. In such cases, the cause of the results would generally be either the gas content or an abnormally high number of suspended particles in the oil. If the gas content is 80,000 ppm or greater, this increases our suspicion that gassing is the real issue. If the gas content is not that high, the appropriate response is to repeat the D1816 determination on another sample immediately, and to also run a particle count distribution using standard method ASTM D6786. This is generally sufficient to establish the cause of the abnormal D1816 results.

If the gas content is high enough (perhaps in the vicinity of 80,000 ppm or greater), there is an increased concern that the gas content has depressed the D1816 value so that the normal, routine testing will not identify the cause of the apparently low dielectric breakdown voltage. Corrective action in such a case is usually not necessary, and would probably be ineffective, in any event. When this situation occurs, it is more appropriate to postpone retesting of the D1816 dielectric breakdown voltage until the next routine sampling interval, and then to include a D6786 particle count distribution analysis at that time. While we will run either of the two gap settings as requested by our customers, our recommendation is that the 2 mm gap should be consistently used. This larger gap is a little less sensitive, so there are fewer cases of essentially "good" oil giving "bad" D1816 results because of dissolved gases.

In the next article, we will continue to discuss the D1816 Standard method, particularly as it applies to new oil, to oil being processed for installation, and to oil installed in new transformers.