Transformer nitrogen leak

[11bgrunt]

I have a station class transformer, 5MVA. In service for three weeks and the nitrogen pressure fell to zero. Gauge needle didn’t go up or down after it hit zero and ambient temp changed. Top oil runs 40C with load fairly constant. Added nitrogen and could not charge past 3psi.Remove feed and pressure would fall to 1.5 and over the next one or two days the gauge would read zero again. This was repeated several times over the next month. Each time the results were the same. I can’t get a clearance on the transformer and the nitrogen blanket is important to me to keep the moisture out of the transformer. During each nitrogen fill, when the pressure was at 3psi, ultrasound suggested the leak was coming from the X1 bushing. Transformer has remained energized during all nitrogen adds.
An external nitrogen regulation system was added that keeps pressure above .5 psi and bleeds off at 5 psi. Since that system was added, the transformer tank pressure has remained positive and the nitrogen bottle pressure has remained constant indicating no gas leak.
Ambient here has been 30-50F.
I cannot explain why the regulation system has worked for the last two weeks and not emptied the nitrogen bottle. Any ideas?

 

[drcampbell]

My first guess is that you have a large bottle and a small leak.
Plug some actual numbers into PV=nRT and see if that provides any insight.

 

[big john]

I agree, the bottle has several hundred cubic feet of nitrogen, and the leak is not moving the entire volume of the headspace, so it will take a while to see it.

Since the leak is slow and you can monitor it, the gas kit works fine. But in cases where we've seen severe leaks we can't shut down to fix we'll install desiccant breathers and just keep an eye on the oil sampling to look for moisture ingress.

 

[iwire]

Hey I do not work on this type of equipment but I can not help thinking the conduits and other conductors are run in a less than spectacular fashion.

Is that normal?

 

[11bgrunt]

Hey I do not work on this type of equipment but I can not help thinking the conduits and other conductors are run in a less than spectacular fashion.

Is that normal?

SO cords are normal and run by transformer manufacture. This installation has a moat for oil containment that does complicate running conduits. This is in a fenced compound that is not accessible to the public. Built in '05, there have been several transformer changes since and we did the best we could under the circumstances.

 

[iwire]

SO cords are normal and run by transformer manufacture. This installation has a moat for oil containment that does complicate running conduits. This is in a fenced compound that is not accessible to the public. Built in '05, there have been several transformer changes since and we did the best we could under the circumstances.

Unsupported PVC supporting LFMC and cables?

No inspections I guess.

 

[gadfly56]

My first guess is that you have a large bottle and a small leak.
Plug some actual numbers into PV=nRT and see if that provides any insight.

I had an earlier post that disappeared for some reason making much the same point. If the last 1.5 psig takes a couple of days to go away it's a low volume leak. That cylinder probably has about 250 ft³ at STP.

 

[Tony S]

I’ve never worked on a nitrogen blanketed transformer so I’d like to know the outcome.

 

[Iron_Bulldog5]

This is just a guess but, if the transformer was initially filled and held under a vacuum it would have lost most of the gasses that the oil had previously absorbed making the oil "dry". Once you added the nitrogen blanket the oil absorbed the nitrogen, each time it absorbed the gas the pressure went down looking like a leak until it reached the maximum saturation point for the gas and then after this point the transformer held it's pressure.

 

[steve66]

I don't work on this type of equipment either, and the previous answers seem like they might have hit the nail on the head.

However, I'll add two more remote possibilities:
1. Are you sure the pressure gauge is working right? I assume you have measured it with another gauge at some point.
2, Or maybe it was leaking from the fill point, and attaching the tank stopped the leak.

 

[11bgrunt]

I don't work on this type of equipment either, and the previous answers seem like they might have hit the nail on the head.

However, I'll add two more remote possibilities:
1. Are you sure the pressure gauge is working right? I assume you have measured it with another gauge at some point.
2, Or maybe it was leaking from the fill point, and attaching the tank stopped the leak.

These possibilities were considered and checked. When filling manually from a large nitrogen bottle, with fill hose connected, pressure would fall when the fill valve was closed and the bottle was turned off. Soapy water was used everywhere possible looking for leaks.