Helium Detectors for Leak Testing

[MechEng176]

Hi folks,

My company does Nitrogen-Helium leak testing as part of pre-commissioning of new oil and gas facilities. Basically we pressure up the system to 95% of the lowest PSV setting with a 99% Nitrogen and 1% Helium mixture and sniff flanges and valves for leaks. The Helium detector gives us a numerical readout of the leak rate - so many cubic feet per year. We currently use Helium detectors made by Agilent/Varian:

http://www.chem.agilent.com/en-US/p...ak-Detection/PHD-4-Sniffer/Pages/default.aspx

these are kind of bulky and I was wondering if anyone had experience with more portable units or a different brand/model they recommend.


Thanks!

 

[BigInch]

You should not be pressuring up to 95% of PSV settings. The real advantage of testing with He is that you don't need much pressure to do it, as large quantities of He leak out through very small holes at relatively low pressures. Otherwise it can be pretty expensive. Find out how to do proper and less dangerous testing, as 95% of PSV settings can be deadly if things went wrong, even with He inside. You will also be able to use less He and reduce your invoice amounts, as less He pressure will be required. Oh, maybe I wasn't supposed to say that.

I hate Windowz 8!!!!

 

[MechEng176]

The piping and vessels on customer sites we work on have already all been hydro-tested. The leak test we do is not to check for strength but if the flanges have been put together properly.

But you're totally right, if a flange hasn't been put together properly and is pressurized to 1,000kPag and someone is standing next to it when the gasket comes flying out that may very well be fatal.

I went over the ASME code and it said for pneumatic testing of piping systems you first pressure to 150kPag or 25% of final test pressure (whichever lower), then slowly work up in increments, stop at those increments to see if the pressure is holding before moving up to the next stage (25% to 50%).

 

[rconner]

I echo all the concerns and comments of BigInch [I did hear however from a fellow now nearly 40 years ago that he was once able to find an elusive leak in a buried pipeline with very low pressure helium and a sniffer that he was not otherwise able to find, as the helium came "up" and earlier water leakage apparently went "down" ;>) ]

 

[MechEng176]

For testing buried pipe with Helium, they generally use 10% Helium to get more to come through the ground. 1% is too low.

 

[3DDave]

I'm impressed with the pictured unit. More impressed that you are looking for something more portable.

I got the chance to use one from Alcatel** that had an impressive performing and impressively named turbobmolecular pump. It would coast down from operating speed over the course of an hour or so. The sensitivity was such that the calibration was from a fist sized 100 psi tank with a tiny orifice and no 'off' valve. Over a few weeks the pressure needle never budged, yet the sniffer could be dialed up to max-out on a reading from that source.

It got creamed when some dofus thought it wasn't working right and jammed the detector hose up to the helium mix tank and gave it a good whiff. I guess a 10,000,000 times higher sample than it was set for was too much. It was repaired, but still.

I think this is it:

http://www.heliumleakdetection.net/...el-asm-110-turbo-cl-helium-leak-detector.html

It's a bit more sensitive, if I'm reading the specs right, but not a rough-field use item.

 

[mbt22]

I've used some of the overall pocket sized ones before. They just bleep with increasing frequency rather than give you a numerical readout, but are great for leak locating. They're reasonably robust, the most common failure has been the sampling tube blocking with bits of lagging. That was usually repairable even then.

Restek

http://www.restek.com/catalog/view/8175

Agilent

http://www.chem.agilent.com/en-US/p...nagement/gas-leak-detector/Pages/default.aspx

Regards,

Matt

 

[MechEng176]

I guess since most pocket sized ones only give an audio signal they won't be good for what we have in mind. When we go to plant sites, each plant tells us their acceptance criteria, eg. a leak of 5ft3/year is a fail. So if we have a unit that gives us a reading of 4 we let the client know but they don't torque the flange but if it's 5 or more, we depressure, they torque the flange, re-pressure, and check the leak again. If it's getting worse then that means the gasket is mostly likely shot and needs to be replaced.

The current Agilent PHD-4s we use can detect leaks down to 0.00Xft3/year. I was looking for a different make/model because as portable and rugged as these are, they are still a bit bulky. But I guess this is the best currently available.

3DDave, I believe your unit is a lab only one. Those are more sensitive to Helium but are a no no when it comes to field testing. Sometimes you'll have a Helium Mass Spectrometer in a test van and an operator will take a hose/probe from it to the flanges they need to check outside and sniff them. That hose can get really long and it just isn't practical when you are leak testing flanges on the top of a vessel.

 

[3DDave]

How does testing in open air result in knowing the flange leak rate? Are the flanges sealed against air infiltration? If you miss capturing the entire helium stream then the leak rate will seem lower than it actually is. Also, how does the helium not stratify in large vessels - leading to an under-report at lower levels and over-report at higher elevation?

Just curious.

 

[BigInch]

He, because it is so mass-less, gets around quite easily and I don't think it tends to stratify quite so much, as does heavier, colder, liquid gases, but I may be wrong.

As far as knowing leak rate, that's not so important as knowing if something leaks or doesn't leak. In most situations that I'm interested in at least, it's a binary thing; pass, or He = fail. No kudos for slow leaks of CH4, or H2S, gasoline, etc.

I hate Windowz 8!!!!

 

[rconner]

I resisted asking the question earlier, but when I heard this was being done with "flanged joints" I also wondered to myself (as I think a great many of these particular are in "exposed"/not buried locations), exactly also how is this particular method would really be more effective (or economical?) than a good soaping of the pipe/joint and trained eye? [I would think in many cases after finding a leak/bubble source, this would in turn be followed by a fix, immediately or at first opportunity.]

 

[3DDave]

We went with the Alcatel in a cold chamber because the bubble solution wasn't cutting it a -55C. The funniest thing was how long it took to figure out that part of the leak that was measured via direct reading flow-meter was just gas getting cold and shrinking in the assembly in the cold chamber. That certainly went a good way to explaining units that 'leaked,' but for which no escaped helium could be found. Turns out if you add a computer to take readings and leave the test equipment alone you can get better test results than if someone is there messing with it to try and 'fix' it.

 

[rconner]

[Notice also interesting post of btrueblood on Valve forum at

http://eng-tips.com/viewthread.cfm?qid=360652

@ 10:14 today.]