Hydro vs Helium leak testing 2

[RiverBend]

I am trying to find a way to compare the leak rate between simple hydro testing vs Helium Leak testing at a level of 1 x10-6 cc sec. I recently had a vessel that was hydro tested per code, 15 min hold time. It was returned due to a leak. Upon retest it took 65 minutes for a single drop of water to form, then pressurized with same psi using 25% helium 75% argon gas and sniff tested at 1 x10-6 and the leak was found instantly.
Need the comparison to justify mandating the helium method.

"Remember, you can't miss fast enough to win"

 

[JLSeagull]

I would limit consideration for mandatory helium leak testing to toxic or other applications with very serious consequences. Helium leak detection is expensive.

 

[RiverBend]

Expensive yes, but not when you consider the costs of having to cut out a leaking vessel out of a cold box and still having to build another and paying shipping costs twice.
I'd be happy to eat the cost of helium to avoid shipping out something that could leak after several hours of pressurization.

"Remember, you can't miss fast enough to win"

 

[deanc]

http://osat-ext.grc.nasa.gov/gso/manual/chapter_07_f5.pdf

Go to the main manual and see 7.20.9. (remove the _f5.pdf)

It is large,but may lead to the info. you want.

 

[weldtek]

Riverbend, I know this doesn't answer the question you posed but keep in mind that your AI may be happy with a 15 minute hydro but he or she may also ignore leaking closures. They're looking for a test that proves the integrity of the design, not necessarily a completely leak tight hydro. Why not hydro test for several hours or overnight?

 

[RiverBend]

I cannot hydro over night or for several hours due to volumn of items & test equipment. I am just trying to prove the higher level of leak tightness integrity the helium test has over the code mandated hydro, which I will still have to do for the code req'd 15 minutes. The AI could care less, all they are interested in is code requirements but I have customer satisfaction in mind and warranty issues.
Thank you for the tips and if anyone knows the actual comparison I'd appreciate it.


"Remember, you can't miss fast enough to win"

 

[moltenmetal]

RiverBend: it is possible for a vessel or piping system to be tight to water and to leak or absorb helium. Capilliary action, permeation etc. are to blame.

We spec helium leakage testing any time we're dealing with hydrogen or seriously toxic materials.

Even a carefully-prepared hydrotest can be insensitive to leakage due to trapped air. Depending on your atmospheric conditions, the leakage water might be evaporating as fast as it's leaking, which also makes it tough to detect. Regardless of service, we always back up a hydro with an air leakage test using shop air at minimum, in case the hydro missed a leak. In fact we usually use shop air at 5-15 psig to find the major leakage before the hydro (i.e. the flanged connection we forgot to torque etc), so we don't make a mess with the hydro water.

 

[JoeTank]

RiverBend,
You don't need anymore info from this group about the technical justification for helium vs water. You already have overwhelming reasons to use helium... it's a business risk issue for you, not a technical one. It' refreshing to see that kind of thinking once in a while.

Joe Tank

 

[ApC2Kp]

riverbend,

You might want to refer to ASME Boiler and Pressure Vessel Section V Article 10 Leak Testing. The helium leak testing sensitivity would be 1 x 10E-06 stc cm3/sec as you have noted. The bubble leak testing method would be much less sensitive at 1 x 10E-02 std cm3/sec. and the hydrotest leak sensitivity would be even less at 1 x 10E-01 std cm3/sec.

The helium leak test is better ({more sensitive) by more than a factor of a thousand times.

It can be difficult to pinpoint the very small leaks. The geometry of the vessel and joints might make a difference in testing with helium pressurization of the part and sniffing for leak of the joint, or a vacuum in the vessel with a small spray of helium on the external surface of joint that is tested.

 

[monkeydog]

RiverBend,

What you are looking for is an old report from JPL, unfortunatly I could not find it for free on the web, you can buy it on the NASA website:

http://ntrs.larc.nasa.gov/search.jsp?R=47466&id=9&qs=No=20&Ne=41&N=4294967278+4294967095

Also, if you are going to perform a Hydrotest, you need to have a dry tank to perform a helium leak test. How about pressurizing with air, and "soap" solution on the welds/joints?

ASM Metals Handbook Vol 11, has a chapter on leak detection methods, that may be of some interest to you.

 

[rzrbk]

As weldtek pointed out above, there seems to be two different requirements being addressed; the hydro test to prove vessel integrity, and a leak check.

It sounds like a low pressure check using helium is the way to go to check for leaks, but your original post indicated that you would be helium testing at full hydro pressure. Unless the vessels and design pressures involved are relatively small, then a full pressure test, using gas, greatly increases the risk to personnel, and test procedures would need to be varied accordingly. (eg. clear area around vessel during test)

 

[monkeydog]

RiverBend,

That link did not work, I the Technical Report is:
"Technical Report No. 32-926" "Basic Criteria and Definitions for Zero Fluid Leakage" by Richard S. Weiner.

A NASA report prepared by JPL dated December 15, 1966

 

[JR97]

I have to agree with what rzrbk posted. Hydrotesting is not solely a form of leak test. Hydrotest is a proof test which verifies structural integrity of the component under stress.

That said, a hydrotest could be considered a pressure change test per Sect V, Art 10, Appendix VI. The sensitivity of a pressure change test is dependant on the volume, pressure, test time, and smallest detectable pressure loss.

My experience has been that using the hydrotest to show leak tightness is not practical. The potential pressure change due to thermal changes during the test time degrades the results to the point that a pressure change would need to be quite large to be attributed to an actual leak versus temperature changes to the water and/or vessel.

JR97

 

[hacksaw]

agreed the hydro is not a leak test but a mechanical integrity issue

the helium leak test is a separate issue all together and not likely performed at the design pressure of the equipment. I'm thinking high pressure equipment here.

During a hydro for example, there may be ancillary devices that have to be removed, thus defeating the leak testing

 

[EdStainless]

RiverBend,
The rule of thumb that we have used is that under ideal conditions hydro reaches 10-2, air underwater or bubble test can reach 10-4, and He 10-6.
One way to look at this is to compare the viscosity of the test fluids, water, air, and He.

The question isn't 'is He more sensitive', it should be 'is He appropriate for my application'.
If this is a high pressure gas HX, or an acid cooler, or in lethal service then the answer is probably yes. If this is in liquid service then it is probably overkill.

= = = = = = = = = = = = = = = = = = = =
Rust never sleeps
Neither should your protection

http://www.trent-tube.com/contact/Tech_Assist.cfm

 

[dcasto]

Hydrotesting piping (and even vessels) under B31.3 is not a mechanical strenght test, its a leak test. You don't even get near the yield stress under the code. B 31.8 will take the pipe to yield at a 72% design factor and thats a mechanical test.

As for leak rate of heliun vs water, helium will leak at a fantastic rate (read money here too) as compared to water. Any gas will for that manner. Use air or nitrogen and you'll get results similat to helium as compared to water. The relationship of helium to nitrogen in leak test will be in proportion to molecular size (close enough).

The next item will be locating the leak. If you have a gas in the system, you'll depend on sound or gas tracers. You could use O2 sensor to find a N2 leak I suppose (looking for O2). You could add a tracer like F6S.

 

[RiverBend]

I am using an Inficon sniffer probe to find the leaks, 25% He & 75% Ni, especially good for areas that are not visible to the human eye.The He testing is definately more expensive but when you look at the big picture, reputation, returns + their shipping charges, and rebuilds, its not really that expensive.
Also ASME section V recognizes it now as an official test method, but it will never replace the hydro for pressure vessel testing, just makes a great (to be sure) test.
Thanks everyone for all the replies.

"Remember, you can't miss fast enough to win"

 

[monkeydog]

Riverbend,
If you are sold on He leak testing, you can cut the He mixure a little more. I have used 15/85 mixture He/N2 with success using a probe. I have seen people use 5/95 mixture. Your leak detector manual, should have a chapter on using a tracer gas.
The plus side of cutting the mixture (besides cost) is if you get a "gusher" you won't overwhelm the background count (too much).

 

[EdStainless]

RiverBend,
I like He leak testing, but you have to know what you are finding. We went to using SF6 because too much of what we found with He was insignificant.
There were also issues with He background levels and He recovery. Yes, we did evacuate test items to recover most of the He.

= = = = = = = = = = = = = = = = = = = =
Rust never sleeps
Neither should your protection

http://www.trent-tube.com/contact/Tech_Assist.cfm