strange noise from transformer, partial discharge?

[JLuc]

Hi all,

I have been asked to investigate a strange noise coming out of three 3MVA 25kv/600v oil immersed transformers.

There is a small noise coming out of the transformer, like a "ting-ting-ting...", about 160 times a minute.

three out of four transformers are doing it. They are all exactly the same transformer.

Is it possible that partial discharge causes this noise? (it seems to come from inside the tank)

We did gas analysis and everything looked fine exept CO and CO2 were a little high.

thanks for your comments.
JL

 

[Zogzog]

Any oil pumps? Would you say it sounds more mechanical or electrical?

Where are you located?

 

[magoo2]

What type of load are they feeding?

Also how long has the noise been going on?

 

[JLuc]

No pumps.

I'm in Canada. The sound is hard to describe. They tried to record it with a digital camera, you can try to hear it, (you need headphones), behind the transformer humm, there's this strange noise about twice per seconds, (it sounded louder on the field).

In case it is partial discharge, what action should be taken. Do you have suggestion of partial discharge tester manufacturer (acoustic test?).

Still in the beginning of investigation , I will update as soon as we get more information.

 

[Zogzog]

I hear what you are talking about, could be PD, but if you had any PD activity you should have increased levels of H2.

PD can be located in oil filled transformers by using HFCT's on the ground combined with magnetically coupled acoustical sensors on the tank. Acurate timing measurements between the 2 sensors are then made and based on the time between the 2 signals and comparing the timing between various locations of the stank sensor the PD location can be triangulated.

ANSI C57.127-2007 has more details on this

My recommendation would be to take DGA samples monthly to track any increase in H2 or other gasses. This one has me curious, keep us updated.

I know a good PD testing company in Canada if you need someone.

 

[ausphil]

It doesn't sound like a PD to me. It definitely sounds more like a mechcanical issue than PD.

Is it there all the time? I agree with zogzog, it sounds sort of like a bearing problem on a fan or oil pump - are there any fans and/or pumps?

The thing is, if you were hearing a PD through the tank wall, you would be seeing huge amounts of gas in your oil samples.

Phil

 

[jghrist]

I haven't heard of acoustic detection of PD under oil, only in GIS or air. Maybe there is PD in the air space. This would explain lack of dissolved gas in oil samples.

What is the construction? Sealed tank with nitrogen blanket or conservator tank?

Is noise at regular intervals or random?

Is noise present with and without fans and/or pumps running?

 

[Zogzog]

It has to be fans or something external to the tank, even with PD in the tank (Which would cause H2 gassin) you couldnt hear it audibly.

Maybe a high side bushing or external terminations?

jgrirst, PD setection in oil is a common practice, see ANSI C57.127-2007 "Guide for the Detection and Location of Acoustic Emmisions from partial discharge in oil immersed power transformers and reactors"

 

[JLuc]

Here are some new infos:

One of the four transformer is a lot louder (one is offline, one is outside so hard to hear).

I attached a copy of the oil analysis that have been made on this transformer.

The H2 is high, 215 ppm. Co and CO2 are also high.

Next step will be a complete inspection to make sure its not mechanical noise, and, complete Doble testing of at least two of the four transformer.

 

[JLuc]

jghrist,

i'm not sure about the transformer construction..

but what I can tell is that the noise is at regular interval and we can hear here it when the fans are not running. (no pumps).

have a look at the DGA , we do see high levels of H2.

Thanks

 

[Zogzog]

Looks like this problem has been present for years, H2, CO, and CO2 have been high all along.

High levels of CO and CO2 indicate cellulous breakdown,(see attachment) which would lead to PD activity, your H2 is also high and has been for awhile.

So what dosent make sense is the H2 levels degreasing over time, however that my be due to sampling errors. I would start with a corona scan of the HV bushings and terminations, dont where else the noise would be audible from, then do a full barrage of testing on the transformer, finally degas the oil.

 

[itsmoked]

I can hear it very clearly too. Not sure I'd call it 160Hz.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[Zogzog]

He said 180 clicks per minute, not second, I counted about 3 clicks per second too.

 

[itsmoked]

You're right it sounds like 3 per second.

Oh, and he said "160" per minute.

Notably when I strip the 60Hz hum out the sound diminishes by the same amount.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[alehman]

Have you looked at the load to see if there is something pulsing at this rate?

 

[Aimn]

I agree with ZOGZOG that this problem has been present for years, H2, CO, and CO2 have been high all along ,and time comes to take action.

Note: May be Gases are generated by normal operation and aging, mostly H2 and CO with some CH4

presents of H2, CO :

-Possible Faults
Thermal fault less than 300 °C in an area close to paper insulation (paper is being heated).

-Possible Findings
Discoloration of paper insulation. Overloading and or
cooling problem. Bad connection in leads or tap changer. Stray current path
and/or stray magnetic flux.

Note:Atmospheric gases (N2, CO2, and O2) can be very valuable in a DGA in revealing a possible leak. Moisture and atmospheric gases will leak inside when the transformer is off and ambient temperature drops.Look for leaks.

Suggestion:
-The weaken point in the Trafo is the tap changer fingers loose ,with time generate a heat and cause gases.open and check.
-check the nature of the load and how is vary or unbalance loads (or overloading)come some times.check if the noise come dowen or not with changing the loads.
-Repeat the DGA for more investigation and comparing results.

I'm interesting to follow this case ,so please keep in touch for any news.

Aimn

 

[itsmoked]

I keep messing with that sound file in my spare time.

I think that noise is just an acoustical result of the three phases humming or beating in some way. ANY filtering, low pass, equalization, etc., ALL reduce the mystery sound just as quickly as the targeted frequencies. How can this be if it is a true ~3Hz source?

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[prc]

Since there is no arcing gases or heat gases in DGA, I would put the reason somewhere else.It can also happen due to loosening of metal parts etc.First try to locate the position from where sound is originating.

Some 35 years back self was faced with a sound problem in a transformer.It was a 12.5 MVA 132 kv Traction supply transformer.From commmissioning day onwards a tuck-tuck sound was coming from the unit at the time of commissioning and also after switching off.It took couple of minutes for the sound to subside.Initially it was reported that sound was coming from all around,but a close watch showed it from one side.When a wooden rod was used as a stethoscope,moving over the tank surface,sound originating point could be located.It was where a tie -rod connecting upper core frame to lower frame was ending.On internal inspection it was found that the locking nut at the end of the tie rod had loosened and fell down.So the long tie rod still fixed at upper yoke frame was acting as a pendulum and hitting on the side of the hole in the lower yoke clamp.This was happening due to the jerk occurring at the time of switching on or off the transformer.

 

[Zogzog]

I just have to laugh, we have a full spectrum analysis post followed by a wooden rod stethoscope post, how far we have come in 35 years.

Funny thing is, seems the wooden rod found the problem. Nice post prc!

I worked with a guy that was a power system engineer in Russia in the 60's, they did "IR Inspections" on terminations by using a candle on the end of a stick, if the wax melted, the connection was hot.

 

[JLuc]

we looked around the txfo and didn't find a specific spot where the sound is coming from.

since DGA does not show significant increase, we decided to take another oil sample in 6 months.

we wil also do a Doble test on 2 of the 4 transformer at the next maintenance.

I believe it is the only thing we can do for now...