As you no doubt know, these are generally attributed as products of cellulose thermal deterioration.
As I recall when CO becomes appreciable fraction of CO2 it suggests decomposition occuring at higher temperature... you have very low ratio so presumably not as high.
I think it is farily common in Generator Stepup transformers to have high CO2 and often blamed on prolonged runs at high temperature.
We had a very similar trend in our 345KV GSU. Oil temperature fairly steady between 80 and 90C. Up to 10,000ppm CO2, 1000ppm CO. Sometimes water bounding above 25 ppm. Oxygen consistently a few thousand ppm (I forget exactly). It existed for 15+ years, during which transformers were drained/dried multiple times.
We assumed temperature was the primary culprit, until we learned different... We replaced our dresser couplings with some kind of hard bolted flange with accordion-type bellows to accomodate thermal expansion.... and the condition disappeared (now CO2 doesn't exceed a few hundred). Apparently it is also the moisture and oxygen (along with temperature) that contribute to this thermal aging of cellulose.
Thank you electricpete in sharing this valuable experience.Can you share a little more details-where exactly you added the bellows and how it stopped air sucking.Was it on the suction side of oil pump?
I am finding excess oxygen in a batch of GSUs in a hydro-electric project.Transformers are water cooled type with an oil pump with water cooler.Oxygen content is around 10,000 ppm and nitrogen 30,000ppm.Co2 only 300 ppm and co60 ppm.H2 2-10ppm and other gases negligibly small.Transformer is with conservator and air cell. Air cell is in tact and so no chance of atmospheric air coming in contact with oil.Can some one throw some light for this excess oxygen content in oil.Situation is continuing for past 6 years since charging.
Presumably the air was leaking in on the suction side because the seal was not good. When we replaced those couplings with a more traditional pipe flange / flexitallic gasket combined with bellows for thermal expansion, we saw 3 changes:
1 - Oxygen dropped dramatically
2 - Water dropped slightly - wasn't tremendously high to being with.
3 - CO and CO2 dropped dramatically and remained low.
Note there was no change in transformer loading. This was a McGraw Edison early 1980's vintage FOA type.
Going back to original question: is high CO2 a problem?
If there is evidence of ongoing generation (comes back after vacuum processing and oil processing), then it is sign of ongoing insulation degradation. How fast is difficult to say. Furan analysis will help judge this same aspect. For large transformers, some people go in and do visual inspection and cut a small sample of insulation from noncritical area for DP test.
