DGA sample confusion

[OffTap2020]

Recently a 15mva transformer at one of our small substations was sampled for DGA.

Our sample showed 6ppm for moisture in oil(D-1533B) and 7ppm for % saturation at top oil temp. The interfacial tension was 39.8


One week BEFORE, an outside testing company, using the same testing lab, sampled the same transformer, and these were there results....

Moisture in oil 17ppm- % saturation at top oil temp.30.1- interacial tension 43.3


Im not understanding this at all. CO2, and O2 were almost equal on both samples.

They were told to retest, this time draining off a larger amount before sampling and the samples came out almost exactly the same.( 16 and 30.2)

A poor sampling method can sure effect results, but how can the results from the same transformer be, in simple terms.....wet, then dry, then wet again? Mind you, all 3 tests were done over a 3 week period of time.

Any thoughts?

OffTap

 

[electricpete]

I would say it is certainly possible that some of the moisture from the oil went into the paper insulation. If I remember right, somewhere around 90% or more of the moisture in a transformer resides in the paper. The partitioning of moisture between oil and paper depends in part on temperature and history. If things are in equilibrium (not necessarily the case), you expect a higher ppm moisture in the oil when temperature goes up, even though the transformer has not exchanged any moisture with the environment. Do you know the temperatures at time of the samples?

Also considering you're dealing with relatively small quantities of moisture, you have to consider the accuracy of analysis and variations introduced by sample handling/contamination etc.



=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[oldfieldguy]

You have to know that a transformer is a happy conglomeration of substances that have an affinity for water that varies with temperature. I fht transformer is cool, the oil has less affinity for water and the moisture will migrate out of the oil over a period of time until equilibrium is established, i.e., both the oil and the cellulosic insulation have absorbed amounts they're happy with for the temperature.

As the temperature climbs, the oil will accept more water, so the water will migrate out of the cellulose, again, assuming sufficient time and steady-state conditions, equilibrium will be reached.

Therefore, if we assume a transformer with a fixed volume of water and oil, the amount of oil in the water will vary over time in relation to operating conditions.

Fun stuff, huh?

old field guy

 

[ELEP]

Did you observed BDV during all samples ?.
If moisture content is increased, it will affect BDV too.

 

[OffTap2020]

We don't as a rule document the ambient temp of the day, while sampling. I guess we should.

Then again, the outside testing company didn't document that as well.

But I'm sure, with some researching, and knowing the dates all samples were taken, I could probably get that info off the net for this area.

This sub has not been in operation for 5 years, but positive pressure has been kept constant, with a automatic nitrogen system, and documentation of good operation.

ELEP....I am unfamiliar with the term BDV..sorry.

Thank you to everyone for there thoughts and comments. I always learn when I come here.

OffTap

 

[oldfieldguy]

Just remember that in addition to temperature, time plays a part. In a constantly changing operating setting, neither the oil nor the cellulose ever really reach equilibrium.

It is also worthy to note whether any of a series of readings is approaching an action level of moisture. If so, you need to dehydrate, regardless of other readings. Of course, readings that are low, low, low, low, HIGH, is quite possibly a bad sampling technique.

I once had a technician go back and resample over twenty transformers that went from low to high in twelve months.

old field guy

 

[ELEP]

BDV-Break down Voltage, to measure dilectric strength.