Transformer oil leaks, How to repair it ? 5

[sparky1976]

Guys,
I have 6 transformers each about 60 MW leaking. all transformer below 5 years old. and all leaks from flanges gasket. I already contact the XFMR manufacturer and no good answer.
I read on some papers that to fix leaks thats not big just use coumpound like glue around the gaskets,
Can anyone advise me with brand, manufacturer or web adresses, adhesives that works on fixing oil leaks ?

thanks
Pitat

 

[bru]

A lot of good hints.

I have repaired many of these.

Here is what I would do if it were my unit.....

Drain the oil, while you fill the Tx with DRY (-55C) air if available. (Maintain positive air pressure) Replace all the gaskets with mild positive air flowing thru windings (still use -55C air!!!) When you want to refill, get a small vacuum pump. The tx can be vacuum filled with a small unit if the correct instrumentation is used. Prior to fill, pull vacuum of 1000 microns or less for 24hr, prefereably during a warm time of year.(obtaining the vacuum with a small pump may take two days). So long as positive pressure from the oil pump is maintained during filling, and the pressure in the tank does not exceed 500 to 1000 microns (ask your Tx rep, if they can even answer, if not, go with 1000microns) the unit may be filled. You must take care that the dew point inside the unit is -30 or lower and that the leak rate is less than say 100 or 150 microns/hr before you fill.



Boil off and leak determination and dew point are possible with the use of remote tank mounted probes. I know of several products that may temporarily fix the leak but they sure make a mess when you go to fix it for real, and you will fix it eventually, that is certain.

Bruce

 

[prc]

What bru suggested is the ideal way of doing it .But in many remote sites in third world it is next to impossible to get all this.Can you suggest some details of remote tank mounted probes for leak detection and dew point meaurement. Model no, make,address or web site?
wbd is suggesting repair of weld leak.Trust he is suggesting to drain oil before welding.I always recommended this rectification in oil filled condition based on some old Westinghouse recommendations,ofcourse with vacuum over the conservator.

 

[bru]

prc
Hastings makes some good vacuum probes, or a simple monometer will do. (this is often the best anyway). You can view their stuff at

http://www.hastings-inst.com/

For dew point, Panametrics in Ireland are well trusted in utility circles. View their stuff at

http://www.panametrics.com/

It may be of value to measure the dew point and temperature of the oil at the same time.

The temperature of the transformer is critical in determining boil off and moisture content. Temperature measurements are best done with a precision milli-ohm bridge on the winding with the appropriate calculations, to give true winding temperature.

The more I think about it, the gaskets are likely of poor quality and have simply compressed. If they have gasket stops, there may be no point in tightening, but if not, the gaskets may have room to compress some more. For poor quality cork neoprene 5 years is about the perfect time for retightening gaskets. This cannot be done with torque specifically, it requires the touch of an experienced transformer man. Torquing will only squish the gasket around and the originator of the torque spec cannot know the density of the gasket that may be in use here.

 

[bru]

Sorry, just a bit more...

These items are not specifically hard to get in north america and you will simply have to add the logistics of getting to the remote areas to thier deliviery. Most of these units are in the posession of utility people and can often be borrowed. I have had good response from suppliers in Canada, in as little as overnight for many products. If you can source 1 clean tanker, one small vacuum pump and 1 small oil pump (300 gal/hr) that should be all you need. Mini oil plants are also not uncommon and will provide both a filter press and vacuum processing with only 600V 3 phase power. The plants will ship on a skid and weigh <2000lbs. Visit

http://www.baronusa.com/

for some examples of these.


The precision bridge may be substituted with a DC current source of approximately 10 amps, a good ammeter and a volt meter. The calculation will show the correct resistance value.

I can post some moisture determination curves and other procedures related to this if you need.

Good luck, Yamin, and email me if you need more info.

 

[prc]

bru,thank you for the information.I believe dew point and temp are not of oil but of air inside tank.If your curves are different from the ANSI curve pl post the same.Dew point meter is the cloud chamber type or different?

 

[bru]

prc..
Dew point can be taken in oil. The probe is designed for it and moisture is derrived from temp and dew point.

I am not familiar with the cloud chamber type. It is simply a probe with matched transducer unit (they are sold together as one unit) and require care in handling but are simple to use.


The air probe and oil probe are each designed for thier specific purpose but the operation is the same.

Oil probes are usually flushed with process oil on partial bypass during process or filling.

I use some GE info and if they are different (I never really compared the two) I will post.
Bru

 

[prc]

bru,
I think we have some communication gap!
In your posting of 27 th Jan you have said of dew point measurement of dry air and limit is-30 c . ANSI C 57.93 also recommends on these lines.Now you are talking of dew point of oil. Why we should check that?
What is meant by boil off?
I have heard GE pioneered these checks to ensure dryness of the paper insulation in trf and also to avoid condensation from out side air.

 

[bru]

prc
The dew point of the oil would be valuable only to check that the oil being pushed into the unit is still dry after being in storage outside the transformer tank. If the oil shows a moisture content higher than 10-12 ppm higher than what came out of the Tx tank originally, you could have problems extracting that moisture as it enters the tank without heat. (We dont have heat if we are not using a processor to put the oil in the tank.

The -30C is the max temp (looking for dew point <-30 if possible) is what we would want to see in the tank prior to filling. This is with the vacuum on after 24 hrs. (Usually 12 is sufficient but this area may be more humid). This may be unattainable if the winding was exposed to humid air for too long without positive pressure from dry air. Theoretically, the longer the winding is exposed to vacuum, the dryer it should get.

It is possible to use nitrogen but the removal of secondary bushings almost allways requires working inside the tank.

Air dryer units can be used. They are made by many suppliers worldwide.

Boil off is also referred to as blank off, and is a measure of how quickly the pressure in the tx tank comes to equilibrium. That is to say, how much moisture comes off the winding insulation before the vacuum in the tank is no longer sufficient to vaporize more water. The tank is blanked off and the pressure in the vessel observed over time. The pressure change vs. time will be non linear for the first few minutes. After that, the linear change in pressure relates to leak rate.

Given temperature and pressure, a moisture content can be calculated. (The Boyles law thing)

Most of the docs I have are from utilties adopted methods but those, as I understand, are from GE, and a bit from Westinghouse.

Just a question for Yamin, if you are still watching, what is the kV of these units, and the BIL?

If they are <=60kV I wouldnt bother with vacuum filling at all but for the fact that the climate is sooo humid.

Also, we must take care that these are suitable for vacuum fill and they were not assembled in a &quot;vapor phase&quot;.

Bruce

 

[sparky1976]

Bruce,
Its a 11kV/150kV Xfmr and the BIL for HV line 650 kV,HV neutral 95 kV, LV line 110 kV.

Guys I already try to glue the leaky flange for 2 days, about 36 hours shut down the Xfmr,During that 36 hrs I try different ways on how to glue it and still leaky.
I use biofix 911 and biodur 561 and it just slow the leak down not fix it, maybe I need to use another glue.

I decide to bring the company/contractor that have xfmr oil purifier on site and change the gasket and top up the oil.

I really want to have viton gasket on my flange, please advise me with company that sell sheets of viton,

Thanks
Pitat

 

[bru]

yamin

you have chosen wisely. I personally reserve patches for equipment near the end of service life, not the beginning. I am sorry to hear that you must encur the expense but it is really for the best.

Some links,

http://www.specseals.com/index.htm

Read about it at

http://www.dupont-dow.com/Products/Viton/viton.asp

Good luck
Bruce

 

[geoffreydavies]

I seem to have a similar problem to yamin. We have a 2MVA oil-filled transformer that has a seam weld leak in the main oil filled steel box part, right underneath where we can only get at it by hoisting up the transformer on a crane.
We are in the Sahara desert and back-to-manufacturer repairs are difficult to say the least. It came from France, along with three others which survived the journey. We have had the weld repaired offsite twice but the metal is thin and the weld seems to open again whenever they re-fill with oil and send the transformer back to us ( 120 miles of bad roads to the local electrical repairer)
Would be pleased to hear from y'all about possible application of external &quot;cold weld&quot; or epoxy product in-situ with transformer full of oil as no-one on site wants to risk taking the oil out due to potential contamination.
Photos available. Thanks.

 

[busbar]

You might consider fastening a conformal plate using the Devcon product. The quanitiy of adhesive needed will probably be much larger, but likely cheaper than a shutdown and oil pumping.

 

[JonBoyz]

Mr. Geoffreydavies

For &quot;coldweld&quot; work I suggest you use Durmetal products. It's far better than any epoxy product in the market. Its easy to use and it works. It even works on oily, greasy and wet surface. Unlike anyother products I've used before.

 

[jbartos]

Suggestion: Visit

http://www.flowserve.com/srvnrepair/pwr_seal.asp

for PWR•Seal™ Technology solutions that will comply with EPA regulations

 

[sparky1976]

Guys,

Probably its little bit late but I think I need to post the result of my repair to leaky xfmr.

Here's what I did,

1. I get permission to spend some money to bring a contractor that specialize on xfmr oil filling with their purifier. We change the gasket on 2 xfmr that the operation give us to have 5 days outage on each xfmr. Change gasket spec from cork nitrile to buna nitrile. it take 12 hor to drain the oil, push by nitrogen, open the leaky flanges let the nitrogen to flow to the air make the pressure of the tank higher by nitrogen (by doing that I use 10 bottle of nitrogen on each xfmr), and change the gasket.
Start to fill the xfmr and purify the oil for 72 hours and let the xmfr sit for 24 hrs and reenergized, result OK, Done that April 2002, so its been a year now and no new leaks.

2. On the other 4 xfmr I only have permission only 1 day shut down, so decide to patch, gave contract to 2 different glue vendor, one Prolong and the other one Belzona, both of them got their technician come to site to apply the glue.
Both almost have the same procedure of applying the glue. Done that in May/June 2002. No laks untill now.

Hope my experince can help some one out there with similar problem.

regards
Pitat

 

[jghrist]

Thanks for the update.

 

[vrnunna]

I think your transfromers are manufacuterd by BHEL. If yes nothing to worry because there is no transformer from BHEL without oil leak.
You can Devcon 5 minutes putty and class A putty for arresting oil leaks. But putty is not a solution for many leaks like valve joints, OLCT joints and tank rim joints.

 

[Bosworth]

I tried the Devcon DurAlloy on a transformer. It worked well at first but did not last very long. After a few weeks the repair cracked on the transformer and oil leaked again. I also did not like that it took an hour to cure. The worked at first and then fell off after a few weeks. The stuff from Velodur Durametal never stayed on. I could not get it to adhere to the oily surface like it said it would. I had to apply adhesive onto a piece of Kevlar and then stick it to the transformer. It was messy and I had a problem applying it before it cured. I did find a product from American Polywater called Power Patch that worked great. It was easy to apply and has been on the transformer almost a year without failing.