Corrosion of Copper and Oxidation of Dielectric Liquids in High Voltage Transformers

Abstract

Breakdowns in high voltage transformers are of major concern. It is therefore a goal to prevent this from happening. Chemical degradation (e.g. oxidation, hydrolysis and corrosion) of the insulation systems and windings and formation of deposits are some of the most important causes of breakdowns in oil-paper insulated transformers.

Several of the methods for studying the oxidation stability of dielectric liquids are time consuming and involve harsh conditions, far from the conditions in the transformer. Some of the methods expose the sample to oxygen flow and uses high temperature. In this work the aim has been to find a quick and suitable method for studying oxidation stability of dielectric liquids under more realistic conditions, with regard to the oxygen and temperature. By using an isothermal microcalorimeter dielectric liquids and model liquids have been studied in terms of the oxidation stability.

This work demonstrates that the isothermal microcalorimeter is well suited for the study of oxidative stabilities of dielectric liquids. The experiments were performed at 120, 110 and 90°C in the isothermal microcalorimeter. The area (integral) below the curves provided by the calorimeter (heat flow vs. time) was found to correlate to the ratio of the volume of the sample over the volume of the head space (amount of available air to the sample). The area below the curves also correlated with the type of oil. It was seen that mineral oil without any additives to stabilise the oil had a much larger area below the curve than the inhibited oil (with additives).

16 unknown dielectric liquids from a Round Robin test were studied with regard to the oxidation stabilities and classified as having high stability, medium, low, very low and extremely low stability by using the area below the curves. Since these samples came from a Round Robin test, the results could be compared with results of investigation of the same unknown liquids by other laboratories that used other methods for determination of the oxidation stability of the dielectric liquids. The results, from the participant laboratories in the Round Robin test that use methods based on the same conditions (observing the onset of the oxidation), agree very well with the results obtained in the isothermal microcalorimeter method, used in this work (based on measuring the integral of the isotherm from the calorimeter).

The isothermal microcalorimeter was also used to study the effect of copper (solid), dibenzyl disulphide and a commercial metal passivator on the oxidation of a hydrocarbon liquid (hexadecane as model oil). This was also done with a vegetable oil and mineral oils. The catalytical effect of copper on the oxidation of the hydrocarbon was observed. The catalytical effect was clear in a base oil with no added antioxidants, passivators or deactivators. The effect of DBDS as antioxidant was also evident in these experiments. The catalytical effect of copper was evident only using a certain amount. Increasing the amount of copper beyond this level removed the catalytical effect. This is sometimes referred to as the critical phenomenon in catalysis.

Also a metal passivator that is used in transformers for preventing corrosion on the copper has been study to better understand the effect of this passivator. Experiments where copper strips were aged in corrosive oil with and without a metal passivator at 150°C over different time showed that the passivator has no effect if added to an aged system where the formation of copper-sulphide has started on the metal surface. The decrease in the concentration of sulphur (measured with potentiometric titration) with time is larger in the samples not added the passivator from the start of the experiment, indicating that the sulphur species reacts with the copper. The effect of adding the passivator before ageing was very good. The effect of the metal passivator Irgamet39 (I39) may be reduced by possible reaction with oxidation products in an aged oil.

Experiments where the passivator was studied were also performed. The change in the concentration (measured with UV-Vis spectrophotometer) of the I39 when heating the solution with the I39 in presence of a copper coupon was measured. The adsorption of the passivator on the copper appeared to increase with increasing temperature. The adsorption of the metal passivator on copper was also studied using a quartz crystal microbalance (QCM) with copper. This gave an indication of the adsorption of a monolayer and multilayers on the copper. The results showed that the adsorption is dependent on the solvent.