Strange Pressure Testing Results

[oelindoe]

Hi all. Not sure if this is the right group.Here it goes:

Im currently involved in pressuretesting of a repairclamp which is mounted over a pipe. During pressuretesting with water at the volume between the clamp and pipe, there is a continous small pressure drop over time but no visible leak... The water volume is 30liters and it is pressurized to 250Bar. After 4 days of testing the leak rate is approx 0.1ml/h. Since there is no visible leaks a hydrogen mix was induced into the water and a sniffer was used to see if there was any leaks. The sniffer didnt catch any leaks! (even at 400Bar!)

So whats causing this pressure drop ?. Are we looking at steel relaxation or continous compression of the water over time. Have anyone experienced such phenomenas before ? Is this a "creep" thing.

All speculations is appriciated since im currently knocking my head against the wall trying to find an explenation...

Brgds Odd Einar

 

[rconner]

250bar (3,625 psi?) sounds like a HUGE external pressure to put on the outside of a pipe -- is the pipe wall sufficiently strong and/or pressurized to be immobile against/withstand such pressure? Until this is established you may not want to pump any more!

 

[zdas04]

30 liters is an amazingly small volume of test water which is bound to be quite unforgiving.

The three things that come to mind are:
1. How are you making certain that the water temperature is absolutely constant (temp change of 1F is about 100 psi)?
2. How are you making sure that the test water is absolutely free of disolved gases (with time even at 400 bar some disolved gas can evolve and give you strange results)?
3. Finally, how are you making certain that you aren't leaking down the pipe (are you testing against a valve or stopple, or using an insert or bolt-on blind)?



David Simpson, PE
MuleShoe Engineering

http://www.muleshoe-eng.com

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

The harder I work, the luckier I seem

 

[oelindoe]

The test performs equally "bad" with and without internal pressure. Buckling is not a problem. The test is conservativly performed up to 400Bar external pressure. The effectice clamp lenght is 400mm with "buckling support" on both sides.

 

[oelindoe]

One last thing i forgot to include:

The results are varying a lot with workshop temperature and all temperatures are logged over 4 days to give a trend curve which shows a slowly dropping pressure.

"ZDAS04" - The clamp is 800mm long mounted over a pipe with both radial and longitudinal seals against the test pipe in both clamp ends.

 

[25362]

A drop of 0.1^oC over 4 days could justify such a volume ("leak") contraction rate.

 

[unclesyd]

I like 25362's post.

SWAG's

Where are you located?
Under the right conditions a very small leak/leaks can evaporate.

How are the two halves of the clamps sealed, IE mechanical or welded?

Did you use H2 or He?
Small amounts of H2 are hard to detect.

What type pressure gauge are you using?

 

[eric2718]

When filling a chamber with gas to around 3000 psi, we get a lot of compression heating. When the filling is complete, and the gas sits in the chamber, the chamber slowly cools back to ambient temperature, therefore the pressure drops with it. When filling scuba diving tanks with oxygen, it is common to set the tanks in a bucket of water during filling, so that they will not be too hot to handle. I imagine this would happen to liquids, but on a much smaller scale, being that liquids are much less compressible.

 

[rconner]

interesting thread --looks like there are a lot of possibilities to perhaps explain what appears to be very small apparent leakage somewhere, though if you have a leak it appears it has not yet been explained how to find? I'm not sure it would show anything, but if it were possible to drain most or all the water out of the interstitial area inside your new clamp if this has not yet been done, and then I wonder if an all air or other gas test even at much lower/safer pressure might show something (e.g. with the same sniffer, or if the entire clamp surface including any seams/joints were "soaped", and/or the clamped area somehow immersed under water etc. to look for fine bubbles)? Of course some gas molecules (like those you have already introduced in some fashion) are quite small compared to water.

 

[BigInch]

If you add a very small amount of gas to that tiny volume of water you will disguise any possible correlation of pressure to temperatue.

It would be nice to see a continuous plot of pressue and temperature through the length of the test.

http://virtualpipeline.spaces.msn.com

 

[oelindoe]

Last update. The "normal" pressure test setup has shown that its not possible to get any decent readings due to the small volume and large pressure variations over time. The test is now completed using a He sniffer and all is fine. The clamp was tested to 450Bar which gave a visible leak due to separation of clamp halves (pressure force > bolt force = correlates to FEA/calc).

Thanks all. At least some useful experiences were achieved