gas leakage test

[rotw]

hello,

I am currently working on the design and testing of an integrally geared centrifugal compressor package. The compressor will compress 100% inert gas (nitrogen) ; however there is an operation mode, which is marginal, whereby the compressed gas will contain amount of hydrogen (enough to classify the gas as flammable mixture).

A gas leakage test is normally performed on the assembled casing to basically test that there is no leakage throughout the casing seal joint. This leakage test is normally performed with inert gas (nitrogen) at max sealing pressure, in my case it is ~ 10 barg. This test is not to be confused with the hydrostatic test and the leakage test of the casing done apart; this leakage test is done on the casing assembled to the gear as ready for shipment.

Here is the problem: the compressor uses labyrinth type shaft seal. So the leakage test appears not doable as the casing cannot be pressurized when the nitrogen will be leaking through the shaft labyrinth seal (on the back of the impeller).

The API 617 recognizes this situation indicating that some leakage may occur with certain shaft seal design ; guess this case is one of those; this seems even more challenging with labyrinth since this is normally a large clearance seal.

My idea is to substitute to Nitrogen, Helium gas and then do the test at reduced pressure in order to decrease the leakage through the shaft seal and get the most out of the pressure which will be able to achieve, in terms of leak detection capability.

Helium has better capability to trace any leak in the casing joint seal for sure; however not sure if this is appropriate and if this will "compensate" for the reduced pressure. In any case, if at all doable, the test can only thought of as special test with the purpose of risk mitigation only, means not substitute of the correct test which is to be done at maximum sealing pressure.

What is your opinion on this matter ? I would appreciate if you can share.

Thank you

 

[monkeydog]

Do web search for JPL Technical Report 32-926
That may provide you with the methodology to perform a alternate leak test.

 

[zeusfaber]

If the reason for doing the test in the first place is hydrogen leakage, then helium (another very searching gas) might be a better analogue than nitrogen anyway.

A.

 

[kirky13]

I think what you are proposing sounds reasonable. Whilst not a true gas tightness test according to API617 in that you probably wont be able to achieve max sealing pressure, it will still be a good check on the integrity of the casing joint.

I have experience with leak testing using a nitrogen/helium mix in complete offshore modules where we couldn't reach the required pressure (110% operating pressure) due to the fact that we had a back-to-back compressor that had constant communication between the LP system (150#) and the HP system (600#) via its balance drum. We could only go to 110% of the LP system for all the piping/equipment on the discharge side of the HP compressor. The test was still considered to be a useful exercise and did give us some comfort on the HP piping. The mitigation measure that we implemented during fist gas was to increase the discharge pressure in small increments whilst testing all the joints.

 

[rotw]

Thanks for your input.

 

[awhicker84]

We blank of everything possible (bearing vent, seal ports, etc) and allow a minimal flow through the coupling laby.

Also, I wouldn't use helium if it wasn't necessary.