Acetylene in transformer insulation oil 1

[riqsoares]

Guys,
We have in our plant a 230 /13,8 kV 31 MVA transformer that had the following DGA result (in ppm):
H2: 228
O2:15400
N2:74000
CH4:41
CO: 297
CO2:1570
C2H4:64
C2H6:9
C2H2:179
Total:818

After this result, we did the ultrasonic test to locate some partial discharge inside the equipment. The result indicated that might have been some problens around the OLTC. So, we decided to call the OLTC manufacturer to do a full inspection at the OLTC.
They found some loose conection between the switch and tap windings. We repaired these connections and we treated the insulating oil to remove the gas dissolved. Then we put the tranformer to operate again. In the latest DGA, we notice that acetylene level is rising again.
My question is: Is it possible that some acetylene was trapped in the transformer paper and insulation and now they are being released or there are another problem inside the transformer?
I am attaching the loose concections picture.

Thanks

 

[physy]

Yes, the gasses are in all of the oil after a time and the oil is held in the paper for a long period even after it is drained. The gasses trapped in the paper will eventually reach an equilibrium with the rest of the oil if no problems are present. When the transformer is refilled under vacuum, some of the gasses will also be removed from the paper. Oil samples should be taken frequently and the results trended to determine if it is leveling off. It may also be a good idea to reduce the transformer loading until a determination is made. Was the oil degassed thoroughly before it was returned to the transformer?

 

[riqsoares]

Yes. The oil was thoroughly degassed before it was returned to the transformer. We did DGA after the job and before load the transformer and the gases level was zero.

 

[physy]

It will take time for the gasses to exchange out of the paper. Have there been any tap changer operations since re-energization? Arcing in the tap changer may be a problem also. Could you post the results of the recent DGA and that of before it was returned to service?

 

[riqsoares]

Yes. There ware tap changer operations after the re-energization. The OLTC is one VACUTAP series.
The most recent DGA: (08/17/2016)
H2: 5
O2:14100
N2:45400
CH4:3
CO:32
CO2:410
C2H4:2
C2H6:6
C2H2:13
Total:61
The first DGA after re-energization (05/07/2016)
H2: 2
O2:4000
N2:13700
CH4:0
CO:0
CO2:230
C2H4:1
C2H6:0
C2H2:0
Total:3
The second DGA after re-energization (05/11/2016)
H2:3
O2:17700
N2:58100
CH4:0
CO:33
CO2:460
C2H4:2
C2H6:1
C2H2:5
Total:44
The third DGA after re-energization (05/31/2016)
H2:5
O2:9600
N2:30900
CH4:0
CO:25
CO2:280
C2H4:1
C2H6:0
C2H2:6
Total:37
The fourth DGA after re-energization (06/29/2016)
H2:6
O2:7700
N2:21600
CH4:0
CO:19
CO2:260
C2H4:1
C2H6:0
C2H2:9
Total:35

 

[physy]

The acetylene level is still pretty low, but the trend is increasing and the elapsed time should be long enough for it to have diffused out of the paper. I would think about a plan though not take any action until the level is a little higher. Definitely increase the frequency of the DGA tests and trend the results, and if convenient, hit it with a round of electrical tests. The increase in the hydrogen and ethane may also indicate that there is something going on, the high temps that make acetylene tend to also drag the lower temp gasses up.

 

[prc]

Answer to original question- Acetylene will not remain in paper and then gradually come out in to oil. Only moisture and furan can remain in paper and then leech out to oil. If acetylene was coming out of paper,other combustible gases also should have come out like that.

Assume that transformer was vacuumed before filling with oil. Current C2H2 level is high (condition 3 as per C57.104-2008) and require prompt additional investigation. Monitor the gas increase. Looks like an arcing (without involving paper) somewhere, may be benign. It can be a loose current carrying contact or equi-potentional connection creating floating potential or core frame grounding at more than one point. Have you checked core grounding? What is the loading level of transformer? Get help from OEM, quite tricky to locate the point of arcing. I had faced quite a few such problems from transformers in service and required long investigation to locate and eliminate the trouble point..

 

[stevenal]

The acetylene doesn't need to be in the paper itself. There will be acetylene remaining in the oil that remains in the paper. This oil missed the degassing process.

 

[prc]

stevenal, electricpete also expressed such an opinion in 2004 in this forum. But my experience for many years is otherwise. Normally transformer is vacuumed for 12-24 hours before oil filling. Then what ever balance oil (including dripped oil) will get completely degassed as these are in the form of thin oil film. Final hot oil circulation will take away further gas. Even if you assume, oil filling is with out vacuuming tank and degassed oil is filled, gas level cannot be to this extent. Dripped and impregnated oil will be only 1-2 % of total oil volume and 290 ppm gas in that spread to 100 % oil, still the gas level cannot be 13 ppm.

 

[stevenal]

OP said nothing about vacuum filling, or any kind of filling at all. My experience with degassing involves a week long hot oil recirculation process performed on an energized transformer serving load. All the gases are nice and low afterward, but rise again afterward; making it difficult to determine whether the initial cause was adequately addressed. I can see that draining and vacuum filling might achieve a better result. Perhaps the OP could state the process used.

 

[physy]

The process does play a role in the removal of gasses. My experience with vacuum refills has been that there is always a shadow of the original left over. It can take some time for the levels to reach a new equilibrium and some may not reach the detection threshold until a problem occurs. OP's 5/7 DGA shows that not all of the gasses were removed as H2, CO2, and C2H4 remained.

 

[riqsoares]

We did 24 hours of vacuum before filling the transformer and we also recirculated 4 times the oil volume, which is 40.000 L, through thermo vaccum equipment.

 

[riqsoares]

Guys,
I have received the latest DGA results. These analysis was made by a different laboratory from the previous ones.

NOV first 2016
(ppm)
H2:9
O2:38200
N2:113500
CH4:3
CO:72
CO2:1030
C2H4:3
C2H6:0
C2H2:32
Total:119

NOV eleven 2016
(ppm)
H2:10
O2:38700
N2:112100
CH4:3
CO:81
CO2:1070
C2H4:3
C2H6:0
C2H2:35
Total:132

 

[electricpete]

I haven't worked with transformers very much since the 1990's. My memory on this subject is fading, so whatever I said in 2004 is more reliable than what I say today. I do recall that certain particular gases were more likely to come back after processing which we attributed as "residual", but I don't remember which gases it was. The one thing I do remember more clearly than anything else is that stevenal and prc are two guys that know their transformers!

I'm pondering now why there may be differences in opinions/experiences. One thing may be the type of transformer and type of processing. My experience was primarily 600MVA 345kv GSU transformers with conservators. They have lots of thick insulation in there so maybe more likely to retain residual oil containing gases (?). And we had various imperfect methods for determining endpoint of processesing such as isolating the transformer and checking the equilibrium pressure/dew point reached. But it was not a perfect system, and there were schedule pressures. We often saw our moisture in oil results shoot back up to 15 or 20 or sometimes 25ppm at about 70C oil temperature within a month of putting the transformer back in service, which is one indicator of incomplete processesing with respect to moisture trapped in the paper. Another indicator: when we finally added a cold trap as part of our processing, we realized that there was still a lot of water coming out in the cold trap long after we would have stopped processing by our older procedures. All this tells me that some of our processing attempts were certainly incomplete with respect to water. Maybe this "incomplete processing" with respect to water might also have some correlation with respect to amount of residual oil and trapped gases? I realize the mechanism for storage within insulation is different for water (absorbed directly in cellulose insulation) and gases (trapped in oil within insulation), but maybe it's plausible anyway? Beats me.

If it is plausible, it might be interesting to know the water results for this transformer before and after and also details about whether it is conservator or nitrogen blanketed.


=====================================
(2B)+(2B)' ?

 

[prc]

In the GSU that electricpete mentioned, 99 % of total water will be in paper and balance only in oil. Typically, in a such a transformer,after 10 years, there can be 198 litres in paper and only 2 litres in oil. Maximum water that oil can carry is 35 ppm at 10C and 500 ppm at 80C and 80,000 ppm in cellulose paper.So when you do vacuum filter oil, you can only remove the 2 litres water only. It will take sometime for the water in paper to come out in to oil and create an equilibrium. This is the reason you are finding the same water ppm in oil a few days,after filteration. So drying out a wet transformer is a long and difficult process.
But cellulose will not trap any gas but only oil -that is my learning so far.The latest results show the problem is continuing and my judgement remains the same.

 

[riqsoares]

One thing that puzzles me it is that the only gas in a raise level is acetylene. Other gases, such as, Hydrogen do not raise its level.

 

[physy]

It looks like there is a continuing issue causing the acetylene formation. A plan for repair should be undertaken as soon as possible. The H2 appears to be increasing also, only very slowly, but H2 production can come from other things. The nature of the problem is likely small and high energy, it may not be making very much of anything else.

 

[unclebob]

Hum, how loose was the connection? This would appear as an hot spot and ethylene (C2H4) should be generated, more than C2H2. To create C2H2, you need a lot more energy and arcing to produce acetylene. This looks like a failed conductor weld.

 

[physy]

I was thinking that it may be intermittent and goes from nothing to a small high energy arc somehow. It really isn't producing C2H2 rapidly.