Twospoons
International Hazard
Posts: 1282
Registered: 26-7-2004
Location: Middle Earth
Member Is Offline
Mood: A trace of hope...
|
|
Oxygen leak detection
I'm looking for a method of finding slow leaks in an oxygen gas dispensing system. The leaks I'm trying to locate are very slow - ml/minute levels.
Soapy water isn't always effective.
So far I've found this : Blue Bottle Experiment which I think could be adapted to leak finding. Relies on redox of Methylene Blue.
Are there any other colour change w oxygen chemicals I could use?
Helicopter: "helico" -> spiral, "pter" -> with wings
|
|
|
j_sum1
Administrator
Posts: 6225
Registered: 4-10-2014
Location: Unmoved
Member Is Offline
Mood: Organised
|
|
Blue bottle is a good idea. But the problem is that it will oxidise in air anyway and so is nt going to selectively change colour for your oxygen
leak. And I think that you will have much the same problem with whatever indicator you use. Interesting problem.
|
|
|
Twospoons
International Hazard
Posts: 1282
Registered: 26-7-2004
Location: Middle Earth
Member Is Offline
Mood: A trace of hope...
|
|
Yes, i'd thought of that too. I could possibly flood the equipment with nitrogen or something else inert, as its all in a plastic box.
Helicopter: "helico" -> spiral, "pter" -> with wings
|
|
|
j_sum1
Administrator
Posts: 6225
Registered: 4-10-2014
Location: Unmoved
Member Is Offline
Mood: Organised
|
|
Maybe make a gel of some sort to smear over possible leak points and cover the gel with plastic wrap to protect from the air. Then when the gel
changes colour you can see where the oxygen came from.
|
|
|
WGTR
National Hazard
Posts: 971
Registered: 29-9-2013
Location: Online
Member Is Offline
Mood: Outline
|
|
Maybe add trace amounts of helium to the process gas, and use a handheld leak tester?
For a typical hand held detector, though, the system might need more than trace amounts of helium to work effectively. If it's in a sealed box, then
perhaps it could be filled with water. Oxygen is somewhat soluble in water, however, so the visual indicator may not be very obvious for tiny leaks.
[Edited on 3-1-2017 by WGTR]
|
|
|
AJKOER
Radically Dubious
Posts: 3026
Registered: 7-5-2011
Member Is Offline
Mood: No Mood
|
|
When one inhales pure O2, the time between breaths shortens. Also, more CO2 is apparently exhaled. Source: see https://www.damninteresting.com/beware-the-dangers-of-oxygen...
This suggests to me potentially a few monitoring tests for a rise in oxygen level in a room.
First, there are monitors to count your breathes per minute as part of a technique to apparently lower average blood pressure. An increase in rate
that is statistical significant could mean a oxygen leak is occurring.
Second, exhaling your breath into lime water becomes increasing more cloudy as CO2 exhalation increases.
Third, there may be a statistical relationship that can be established for a person frequently present in a room where a potential leak could occur
with respect to heart rate, for which, there are many monitors available for the like of runners.
Fourth, capture air in a glass test bottle and spray a predetermined amount of say acetone into the bottle and quickly light and record via a ruler
standing by the vessel, the height of the flame and/or burn duration. If the fuel is supplied in excess to normal O2 content in air, an increase in O2
may be visible by the flame height or burn time. Similarly, use H2 as the fuel supplied in excess, and record the loudness of the retort. I like this
last test due to the cube rule where doubling the amount of an explosive approximately increases the power of the explosion by 2 cubed or 2x2x2 = 8.
This scale effect may be discernible in sound and allow the detection of say 10% increase in oxygen concentration via an explosion that is now over
33% stronger (1.1 x 1.1 x 1.1 =1.331). In place of sound, one could measure the distant that a cork stopper is shot at a 45 degree angle.
Whatever the measure selected (call it the 'y' variable), one would construct a graph (and confidence interval bands) based on actual oxygen content
(the 'x' variable). The analysis then becomes an inverse regression point estimate based on a given observed y (or better, the average of repeated y
observations) and the likely associated x and its error interval. The standard deviation of the error measures follows by assuming a Bivariate Normal
distribution for x and y, and then noting the conditional distribution of a normal variable 'x' (now, no longer a known fixed constant as in classic
regression) given 'y' is normally distributed with values determined by theory (see https://onlinecourses.science.psu.edu/stat414/node/118 ).
The accuracy of all the above suggestions will most likely need field testing.
[Edited on 6-3-2017 by AJKOER]
|
|
|
Sulaiman
International Hazard
Posts: 3558
Registered: 8-2-2015
Location: 3rd rock from the sun
Member Is Offline
|
|
put sometthing really smelly in the oxygen feed and use olefactory sensing.
e.g. mercaptans https://en.wikipedia.org/wiki/Methanethiol
that is how natural gas leaks are detected here in UK I believe.
If you can shut the system down then a vacuum test could give lots of useful data.
(overall seal integrity, leakage rate, entrapped water ...)
wrap piping in iron wool and look for heat ? 
A thermal imaging camera ? (with or without the iron wool)
(cooling due to expansion)
[Edited on 6-3-2017 by Sulaiman]
CAUTION : Hobby Chemist, not Professional or even Amateur
|
|
|
AJKOER
Radically Dubious
Posts: 3026
Registered: 7-5-2011
Member Is Offline
Mood: No Mood
|
|
A variation of Sulaiman's suggestion would be to first move the O2 tank outside. Then, add a mercaptan to Vaseline and place on all seals. Further
coat the doped Vaseline with pure Vaseline eliminating odor. Assuming any escaping O2 gas will push thru the layers, the smell should eventually
escape and be detected.
Issue with this idea is the seals would likely to have to be retreated with time.
There is also no metric to indicate the magnitude of a problem.
|
|
|
![[*]](images/xpblue/default_icon.gif){kind=icon}
