Need advice from someone with real world transformer DGA experience.

Status
Not open for further replies.
11bgrunt

11bgrunt

Pragmatist
Location
TEXAS
Occupation
Electric Utility Reliability Coordinator
Transformer DGA experience.
3750kVA oil filled transformer. 24900Y/14400 high side, 480Y/277 secondary.
DGA in '09 and 2010. H2 was 3000.
Recent test shows H2 at 9000ppm.
Transformer is running at 90% 24/7. Oil test has everyone excited.
Does anyone have field experience that would force me to get my spare ready and schedule a change out as an emergency? I haven't seen enough of the report to see the ratios of the other gases. The Hydrogen is what prompted the recommendation to do something quickly.

From the WWW I found a a few lines that discuss DGA.

Copyright © 2008 DSI Ventures, Inc. All rights reserved.
DGA interpretation is not an exact science, as there is a lack of positive correlation between laboratory data and field experience.

Dissolved Gas Analysis (DGA) diagnostic methods. Much progress has been made in the last decade
in understanding the relationships of gases in transformer insulating oil and diagnostic
outcomes have become more accurate. Yet field practice doesn’t always reflect the newer insights.
 
I have seen this on wind farms. The cyclic nature is a different phenomenon than your steady situation. We saw increases in the H2 and it was not cause for a knee jerk reaction.

When was the trf installed/put in service?

Do you only have two data points(tests)?

What lab did you use or was it a field test?

The hot metal gasses are more concerning for immediate action.
 
MD84 said:
I have seen this on wind farms. The cyclic nature is a different phenomenon than your steady situation. We saw increases in the H2 and it was not cause for a knee jerk reaction.

When was the trf installed/put in service?

Do you only have two data points(tests)?

What lab did you use or was it a field test?

The hot metal gasses are more concerning for immediate action.
Click to expand...

This was a new transformer when installed 4th quarter '09 after ex transformer failed. DGAs for three consecutive quarters after install. Lab work done by Doble.
A second DGA this week showed H2 increased at 50ppm per day. This is forcing transformer change out tomorrow.
 
There's nothing imprecise about gassing that severe: Even 3000 PPM is a massive amount of hydrogen and should have been a flag to take immediate action.

Hydrogen doesn't develop much just from overload, but typical values are in the double-digits, with PPMs in the hundreds being cause for concern.

At a bare minimum that transformer should've been power-factor tested after that first sample. And depending on the the results, the oil could have been processed in the unlikely event the insulation looked good.

The ratios of othet gases are just as important, but at a minimum a jump from 3000 to 9000 indicates major electrical discharge, so I agree that transformer needs to come out of service immediately.

What did Doble say? They usually aren't shy about speaking up over serious problems.
 
50ppm/day seems like an active fault. The condition we were seeing was anywhere up to 5000ppm H2 in the first year of operation. Subsequent testing showed this leveling off. We were not seeing anywhere near 50ppm/day.

I would be curious to see what other gasses are being generated with the H2.

If you are planning a RCA of the existing unit and care to share results please do.

A power factor, excitation, and ttr could be reveiling.
 
MD84 said:
...The condition we were seeing was anywhere up to 5000ppm H2 in the first year of operation. Subsequent testing showed this leveling off....
Click to expand...
What other gases were you seeing in addition to the hydrogen? Were you cycling tap changers?

I have literally never seen a transformer with hydrogen content like that which wasn't actively failing.
 
No other gasses in considerable quantities. Some methane, ethane, and CO iirc. It has been a while since I have been working on that job. The wind farm transformer is a different animal. Have you worked with these? They have a very dynamic load profile.
 
MD84 said:
No other gasses in considerable quantities. Some methane, ethane, and CO iirc. It has been a while since I have been working on that job. The wind farm transformer is a different animal. Have you worked with these? They have a very dynamic load profile.
Click to expand...
I know they can be peculiar, we were testing brand-new oil GSUs with power factors in the neighborhood of 10%. We actually had to put in a call to the OEM who said somehow that was perfectly normal.

I'd have to pull up DGAs, but I'm not aware of them having unusual gassing. Maybe they do.
 
I had some cooper's with FR3 that were perfectly normal. The ones with mineral oil exhibited this issue. They were GE Prolec. The initial operational DGA was performed several months after being put into service. Some exhibited these high TDCG levels and were immediately put on a monthly DGA and weekly TCG schedule. The generation rate was low and the units did not increase over time so they were put to an annual cycle.
 
Everyone agreed that transformer had to go. The concern was an explosion as the combustible gases exceeded 10,000ppm from the DGA.
The new transformer is in and the plant is back in service. I have a RF sniffer with an ultrasonic option that is used on RFI complaints, usually from HAM operators. I have studied a little on ultrasonic detection for MV gear and took the unit I have to run test on the transformer being replaced. Secondary was over 3000 amps when a sweep of the old transformer wall was done. Noise was detected on the tank wall under the secondary bushings only and always came back to the same spot as the tester was moved around the tank.
The new transformer was installed and before any real load was applied, the ultrasound test was done. The entire transformer was quiet with no noise detected.
Twelve hours later and the plant is back over 3000 amps on the new transformer.
Ultrasound detected noise on the wall under the X2 and X3 bushing like before on the old pot but now we pick up noise on the wall under the H3 bushing.
We have similar transformers at other plants but they do not run with this type of 80-90% continuous load. Tomorrow I will go to those plants and sweep those transformers with this small ultrasound for some type of reference.
Transformers and labor are expensive when you can schedule a clearance. Everything gets worse when it just stops working in the middle of the night and out of warranty.
Does anyone believe this is a waste of time?
Five years ago I asked a testing company how they felt about partial discharge testing and ultrasound detection.
Their reply was "that is black magic hooie". They still do not offer those services today. I have looked at PD testing equipment but never bought anything.
I plan to be involved with the transformer rebuilder as they they go through their testing to determine what is wrong inside. They have told me the first test they will do on the old transformer is PD.
Many people have said the most common point of arcing is the no load tap changer that never gets moved and is the only non bolted connection in the transformer. I am recommending that a NLTC be removed from our transformer spec sheet.
 
PD testing is a very good test. There is no question in my mind of its validity. Online PD testing is not nearly as exacting as a factory grade offline PD test.

If you are concerned with these transformer installations I would suggest developing a procedure for commissioning, predictive, and preventive maintenance. The PD is one to include in the commissioning. Not sure if you already have something place. Are you currently performing any commissioning testing on newly installed units?

Many times it seems in depth testing is reserved for larger transformers however evaluating how critical this piece of equipment is could warrant this testing. There are many variables to consider such as lead time for replacement or cost of spare as well as lost revenue during downtime.
 
Commission test on accounts outside the substation fence are voltage and rotation checks.
If a primary fuse doesn't blow them you are golden.
On site PD by a service provider is around $3K a day. We do thermal inspections and on request we will order a DGA like this one account.
I see so much of this testing is almost subjective and dependent on the tester and reviewer of the data.
 
Status
Not open for further replies.
Top