4
$\begingroup$

I'm trying to increase the Helium detection capabilities of a sniffer unit. Could a cyclonic separator, like that used in the oil/gas industry, with enough fluid velocity move the heavier gas molecules in air (O2, N2, Ar, CO2) to the sides of it while keeping the lighter elements (He, H2) in the center to be detected by a sniffer?

We're trying to detect leaks of helium from refrigeration systems on an assembly line. we're designing a positive air box to prevent any outside contamination from the filling and recovery stations. The box will have a blower attached to it to pull the volume of the box in 10 seconds. because most of the gases in the box are not our trace gas we have a very low mbar l/s level (allowable is 9e-7) that is on the borderline of the minimum of our sniffer unit we are looking to use. The sniffer unit has the precision of 1e-7 mbar l/s at 300 SCCM and 1e-6 mbar at 3000 SCCM. We are going to have an array of sniffer heads in the center of this conic that will sample the air at the center (with the thought that the lighter elements would be there) leading to a more accurate reading because it is sampling less of the non-trace gas. This process of sampling and detection along with the lifting/closing of doors and shifting of the equipment is done on a conveyor belt within a 40 second time period.

The current system is a person at the assembly line with a hand held sniffer going around the connections, but are getting readings that there are leaks even though there might not be, probably due to the helium charging station only 3 feet away from the leak testing area. we are looking to automate the process for less false readings and quicker turn over time in the process.

Improve this question
$\endgroup$
1
  • 1
    $\begingroup$ @Air i added more detail in the actual question, but i am not sure as to the accuracy of the current setup because this is a customers setup we are trying to modify. However the main thing is we are looking for more accurate quality control for the process. $\endgroup$
    – Starky_95
    Dec 28 '16 at 18:24
1
$\begingroup$

Cyclones are designed to quickly and cheaply remove contaminants that would tend to settle out of your process gas (or liquid) if it stopped flowing. They're very good at controlling coarse PM, not so good at controlling fine PM and really quite poor at controlling ultrafine PM. What you're looking for here is so far past ultrafine PM control on this spectrum that you would never achieve a useful result.

Even if you could slightly concentrate your heavier gases around the outside of the cyclone, where are they going to go? In a cyclone, the volumetric flow is the same at the inlet as it is at the outlet. The only thing that gets knocked out is the particulate, when it impacts the wall and settles down into the hopper. This works because that particle would tend to settle anyway, if it weren't entrained in your process flow, and near the wall of the cyclone the flow approaches zero (recall the no-slip condition of fluids 101). The same can't be said of a theoretical packet of slightly-heavier-than-average air.

We are going to have an array of sniffer heads in the center of this conic that will sample the air at the center (with the thought that the lighter elements would be there) leading to a more accurate reading because it is sampling less of the non-trace gas.

I'm not sure this really ends up being more accurate, given that your sample is no longer representative of the mixture on which your leak detection threshold of 9 x 10-7 mbar·L/s is based. This isn't a situation where you want to set your leak detection threshold at zero.

I think any clever gas stratification scheme would just over-complicate the process. Using the PPV box might be enough by itself to make your client happy. Otherwise, if you just need a little bit more sensitivity and you can't go with a more sensitive device, I'd keep thinking about optimizing the sampling conditions—can your client afford to spend 90 seconds per unit rather than 40, if it cuts down on false positives? Or maybe your client would be willing to perform some more reliable leak test on a sample, rather than testing every unit, if they can accept a non-zero defect rate.

Improve this answer
$\endgroup$
2
  • $\begingroup$ I can ask my supervisor if we can increase the time, and we were going to see about using a multitude (6 or7) of sniffers at 300 SCCM rather than one to help and try to increase the sensitivity as well. we can pose it to the customer and see if we can get a time increase, because from what i understand thats the time limit on the conveyor belt (or at least how my supervisor put it) $\endgroup$
    – Starky_95
    Dec 29 '16 at 14:29
  • 1
    $\begingroup$ Their line runs at a certain speed and you'd be slowing that down. I only suggest it in case you can actually get a better reading given more time. It may be that 40 seconds is enough to get as good a reading as you can, per sniffer. It may be that six sniffers of the type currently in use is affordable, but one sniffer of a better type is not. Those questions are worth investigating. Hand-held units aren't always the most reliable. $\endgroup$
    – Air
    Dec 29 '16 at 16:29
0
$\begingroup$

Gas centrifuges (I assume that's what you mean by cyclonic separators) certainly work. They are used to separate Uranium-hexaflouride with U235 and U238. The relative densities of helium and components of air is much greater.

Getting good information on gas centrifuges may be tricky because of their military use, and digging too hard may put you on some watch-list.

Improve this answer
$\endgroup$
4
  • 1
    $\begingroup$ What i saw on google is not what i'm looking for, the design is completely different with the first picture that popped up on google. The separators that i'm looking at are solid-gas seperators typically used in dust collection systems and flue gas recovery to remove the soot before it leaves the chimney. would the principles used in the solid-gas separators also work with a gas-gas mixture of trace helium and atmospheric air work? $\endgroup$
    – Starky_95
    Dec 28 '16 at 21:09
  • 1
    $\begingroup$ @Sta: Sorry you didn't like the answer, but your question was rather ambiguous. In any case, a simple gas centrifuge might be exactly what you want. Apparently you're talking about something that uses inertia of the gasses to swirl them. Ultimately, you're still separating by density. It probably comes down to the concentration ratios you want. $\endgroup$
    – Olin Lathrop
    Dec 28 '16 at 22:26
  • $\begingroup$ it's okay, i saw those when i was initially started doing my research and found that's not what we needed so i shifted my search to things like the inertia based ones found in the oil and gas industry. $\endgroup$
    – Starky_95
    Dec 29 '16 at 13:58
  • $\begingroup$ @OlinLathrop, this was my first thought as well. "A centrifuge can do it because that's how they enrich uranium." Main difference between a centrifuge and cyclonic separator is that the centrifuge has some sort of agitator or rotor inside to effect the outward force. A cyclonic separator uses a tangential inlet and the motive force of the process gas. End of the day, they're both relying on differences in density. $\endgroup$
    – Byron Wall
    Jan 4 '17 at 23:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.