Water content in oil

[Gauss2k]

We regularely take oil samples from our clients transformers in order to analyse their dissolved gas content and oil properties.

Sometimes (and it happens a couple of times in the last few months), we see a sudden change in water content (for example from 10 ppm to about 50 ppm a year later), without any major difference in other parameters (dielectric strength, etc). When it happens, we go back and take another sample in order to confirm the increase and, most of the times, it's much lower and thus closer to the last year sample.

Of course, I understand that a sample can be contaminated by water or humidity (to minimize this, we take at least 1 to 2 liters of oil and operate the valve in order to remove all the trapped water before taking the sample). This would be the main reason for a higher water content.

But is it possible that the higher water content be the real value? Would it be possible that an oil sample taken from a transformer with an global average water content of 50 ppm be at 10 or 15 ppm? I was not able to find much informations about how the dissolved water content is distributed inside a transformer. Can some oil water content be significantly higher or lower than the average?

 

[isquaredr]

Hi Gauss2k
Try to avoid taking samples in wet conditions and always clean the sample bottles twice with oil taken from the sample point before taking the actual sample.
This helps to clear moisture both from the sample bottle and the sample point.

 

[Gauss2k]

We also already do that, we clean the jar and syringe with the transformer's oil three times, after removing the 1-2 liters We take a lot of precautions to avoid contamination when taking the samples. That's why I wondered if the higher result *could* be real in some cases..

 

[oldfieldguy]

Let's assume a fixed amount of water inside your transformer. If so, there is an equilibrium reached for a given temperature. Hot oil will have a greater affinity for moisture than cold oil, so with each changing set of circumstances, the moisture content of the oil will change.

Changing temperature conditions will take some period of time before the moisture reaches equilibrium at the new temperature.

All this being said, the measure of moisture in the oil will change without changing the total amount of water in the transformer.

If you're concerned about the numbers, have your transformer "dried".

old field guy

 

[unclebob]

Can you establish a correlation between the water level and the sample temperature?

 

[oldfieldguy]

unclebob--

Nope, not on my watch. You have the cellulosic insulation of the transformer. It has an affinity for moisture. You also have the oil, which has its own affinty. When the transformer is "cold" (a relative term) the moisture will slowly migrate from the oil to the cellulose insulation.

As the transformer heats up, the process will reverse.

To compund the issue, if the temperature transition is rapid and there is a greater level of migration OUT of the cellulose than the oil can accept, moisture will precipitate out of "solution" and be visible as a foggy appearance to the otherwise clear or yellowish oil, and may in some cases drop out of solution entirely and end up in the bottom of the transformer.

Free water in your transformer oil sample simply means that under the circumstances of time and temperature, you have more water in the tank than the oil and the cellulose can accept. You can have considerable fre water today and almost none tomorrow, depending on temperature and time.

However, if at any time you find free water in your tranformer oil sample, or you have high moisture content in your tests, know that your transformer is sick and in need of attention if you wish a long and reliable service from it.

old field guy