Transformer DGA Results 2

[mbk2k3]

Hey folks,

So I've been reading lots of presentations/IEEE papers trying to find out the root cause of a transformer failure.

The brand new transformer was commissioned less than a year ago, all protection trip signals for high temperature and low oil levels were wired to a protection relay and commissioned. the temperature and oil level alarms were also wired into a PLC and they don't show any change in state either.

I'm attaching the DGA results screenshot here.

What confuses me is this:
[ul]
[li]The Ratio Tests (Rogers Method/Basic Ratio Method) point to High Energy Arcing Discharge.[/li]
[li]The Key Gas Method points to Electrical High Energy discharge as well, but potentially also a "thermal oil" fault type as well.[/li]
[/ul]

Can a high energy arcing discharge occurr randomly out of nowhere without being preceded by a thermal type overload? Or was there a long term overtemperature condition as well that ended up resulting in an internal arc? The methane levels are really low, and I believe if there was an overloading condition the methane levels would have been higher....

I checked the metering data and there doesn't seem to be any overloading condition in the past few months or even the day the failure occured.

 

[RRaghunath]

DGA results indicate high level of Acetylene. This is clearly an arcing fault.
The transformer needs an internal inspection to find out where the arcing fault was there and further to rectify.

Rompicherla Raghunath

 

[prc]

It looks a strange pattern. Please retest on a new oil sample. Is there an OLTC? Before opening the transformer check whether there is any leak from Tap changer chamber to transformer tank.

 

[mbk2k3]

It looks a strange pattern. Please retest on a new oil sample. Is there an OLTC? Before opening the transformer check whether there is any leak from Tap changer chamber to transformer tank.

there is a tap changer, not sure if on-load, but i'm certain its never been used (if that makes a difference).

what are you suspecting?
and what do you think makes it look like a strange pattern? do you typically see gas numbers that tilt 100% in one direction (thermal failure vs electrical failure) as opposed to a mix like what we have?

 

[prc]

There is a big tilt in acetylene without the presence of other thermal gases and hydrogen to the same degree. Oil in OLTC chamber will have high content of acetylene and if it mixes with main oil, such condition can be seen.

 

[unclebob]

The lifespan of a transformer goes from cycles to centuries...

This sample is after the fault occurred? Without a pre-fault DGA, it's very hard to tell why it happened.
How is the condition of the oil?

 

[krisys]

Are these test results of DGA done after some fault trip of the transformer or is it a routine test?

If it is a routine test, you may conduct another test after a month or two and compare. If the trend is giving some clue, then act upon.

Just to clean the slate, you have to filter the oil (including the de-gassing). At that time perform the DGA and keep it for the reference. Another reading after two months, would tell you if there is any concern.

 

[prc]

The procedure depends on the test results. An acetylene value of 42 ppm is quite high and in most of the cases, situation will be grave for an immediate shut down. In such a condition there is no need for degassing or waiting months to see the increase in gas content, by that time a catastrophic failure will be the outcome. I doubt the test result. So a retest,preferably in a different lab. I prefer to do DGA analysis not by ratios or Duval triangle, but by pattern analysis. I have a data bank where I keep all gas patterns(taking max fault gas as 1.0,put others in per unit basis) versus actual fault found inside transformers. I find this much easy and fast for work.

 

[mbk2k3]

I have a data bank where I keep all gas patterns(taking max fault gas as 1.0,put others in per unit basis) versus actual fault found inside transformers.

Hi PRC! would you mind sharing this data? It would be really helpful. Perhaps a dropbox link?

Additionally, I can show you pictures of the fault we are dealing with.

Are these test results of DGA done after some fault trip of the transformer or is it a routine test?

sorry i should have been more explicit in the OP, this was a catastrophic failure on the transformer. it didn't blow up, but internally one of the windings got shattered and lots of bent copper/metal.

 

[krisys]

First of all it should be clear in our mind why we are going to conduct the DGA?
It is basically a diagnostic test.

For a transformer in service, this test gives the clue, that what is happening inside the transformer. Internally the transformer is like a chemical factory. Based on the transformer condition, lot of changes taking place. The gas released during this process, gives a good clue, whether transformer has some in-detectable sparking, discharges or heating up due to some reasons such as overload etc.
It tells you how urgently you have to attend the problem, if it exists.

In your case you know the transformer is already faulty and needs attention. So the DGA would give some clue to the root cause analysis (RAC). But, it may not help much.

When the patient is dead, why you need to conduct the diagnostic tests on the body?
Only the post mortem will be conducted. (Only on a lighter vein, no offense please.)

 

[prc]

krisys, You have aptly clarified. But DGA will not give any clue after a violent failure for RCA. In fact it is meaningless to do such a test after a catastrophic failure. Most probably this test seems to have been done after a couple of days after failure. Any way my thinking was correct -results are strange.