Leaks Allowing Water into Compressed Air Lines 4

[tlee123]

In a recent thread (thread378-122965), someone presented anecdotal evidence that leaks in compressed air lines allow humid air from the factory to enter the air lines. This of course caused a problem with wet air downstream of the leaks.

Can this possibly be true?

How can the low pressure humid air move against the high pressure gradient into the pipe?

Any ideas? This has been bugging me since I read it.

Thanks,

Tom

 

[jcfoley]

Tom,
I used to work for a major industrial gas supplier. This is true. I worked with high pressure helium and any leak will cause an impurity increase. On the theory side I believe that it involves both vortex theory and statistics.

The flow around a leak can create localized pockets of reduced pressure. Modern statistical flow theory says that any molecule has a small but finite probability of moving against the pressure gradient.

With that being said, I work where the rubber meets the road, not in some text book. But, take a helium cylinder with a known impurity analysis (99.9999% pure Helium). Open the valve to atmosphere and vent it for say 10-15 seconds. Then, close the clyinder valve. The impurity analysis will show a higher level of air contaminates in the pure gas.

I do agree that this does not sound right , but I have solved way too many process problems to not believe it.

Chris Foley
Midland, TX

 

[unclesyd]

There is a very old paper from Gulf Research (The Old Gulf Oil Research Group) that documents the leaking of O2 into a He system operating at 22,000 psig. The source of the leak was the packing of a valve. If my memory serves me correctly the packing was leather. The lab was instrumental in development of the packing less valve to prevent such an occurrence. Also they have been several articles on this problem published in the industrial heating arena.

The Oxygen diffusion constant is very high somewhat close to that of Helium.

We use very high purity N2 in our process that is supplied in headers at 125 psig. To keep the O2 out this system there are very few flanges and valves. The letdown valves are the biggest culprit in allowing O2 ingress. It's a constant chore on the low pressure (3 psig) system keeping the packing and flanges tight on the remaining packing type valves.

Here is one type of an O2 diffusion leak demonstrated . The introduction of the low concentration O2 analyzers really brought the problem of O2 diffusion to light.

http://www.a2c2.com/articles.asp?pid=323

 

[tlee123]

Thanks Chris and unclesyd.

So it appears true that leaks do contaminate high purity gas and the contamination mechanism is diffusion against the pressure gradient.

However, I'm still not convinced that a significant amount humid air could diffuse through leaks in a typical compressed air lines. By "significant", I mean enough contamination to cause liquid water to form downstream of the leaks.

I suppose a good test would measure the difference in the dew point of the air upstream and downstream of some intentional leaks.

Tom

 

[jcfoley]

Tom,

I would urge you to consider a valve packing with rain pooling on the top of it. I have found liquid water in systems with bad valve stem packing and the dewpoint was supposed to be -40F. Yes, these were realy bad cases but it can happen.

Chris

 

[swainw]

Partial Pressures. Remember Boyle's law of partial pressures and the math will show how this occurs. Also, the transfer theory associated with molecular gradients of an constituent drive the entropy of a system to move towards equaliztion. High humidity to low.

The air pressure is a molecule in the airstream that fights the outside air entry, but the partial pressure of water vapor will cross the boundry to become absorbed into the free airstream as it has a low partial pressure of water vapor.

Bill

Bill Swain
Ultra Electronics Precison Air Systems
Technical Coordination Manager-US Programs
swainw@asme.org

 

[rmw]

tlee123,

Difussion, which I first encountered professionally with zirconium insitu type oxygen analyzers on boiler outlets, initally was very a difficult concept for me as a mechanical engineer to grasp. We did not study that kind of stuff in the course I took. We studied the kinds of flows that you could measure with instrumentation. Then, some sharp salesman used an example that illustrated it perfectly for me.

He said "have you ever been sitting in a large auditorium, a church for example, and some one passes gas (farts) at some distance away from you, with absolutely no ventilation or other type of air movement in the room of any sort at the time, and within seconds you can smell it? That is difussion."

I grasped it immediately, and have understiood it ever since. Gratefully, because of difussion, we don't suffocate in the CO2 we expell in still rooms.

Old Mechanical Engineers can be taught new tricks.

rmw

PS, in the thread you referred to, you have to remember that the compressed air piping that ran from the power house of the paper mill to a newer paper machine some distance across the plant site, ran through building(s) of other older paper machines, which is a real humid environment, irrespectdive of the ambient humidity level. Lots of steam vapor, humidity from the paper drying process, and steaming liquids around this area.

 

[olynyk]

Swainw is onto what I believe is the culprit:

If the compressed air is very dry, even though it may be at a high pressure, its partial pressure of water vapour will be very low (essentially zero). The pressure of the outside air may be much lower than that of the air inside the pipe, but the partial pressure of water may actually be higher. In that case, the water will diffuse against the pressure gradient due to the driving force of entropy.

It's the same thing that makes a living cell swell up and explode when you put it into pure water (instead of slightly salty water) -- the water diffuses against the pressure gradient into the cell.

 

[tlee123]

Thanks Bill Swain for bringing this dying thread back to life!

I did some more research on partial pressures and diffusion (BTW it's Dalton's Law of partial pressures, not Boyle's).

So I agree that the partial pressure of water vapor outside the pipe is higher that the partial pressure of water vapor inside the pipe. Therefore, with a small hole in the pipe, water vapor will tend to diffuse into the hole.

Regarding olynyk's example of liquid water diffusion into a living cell:

Yes, the water diffusion (osmosis) will occur even against a small pressure gradient. However, when the pressure gradient rises, the diffusion direction reverses. This reverse osmosis is a technique for purifying water.

This brings me to my point. I think of the diffusion of water vapor into the pipe like salmon trying to swim upstream. They can do it, but only if the flow of the river faster than the speed they can swim, right?

The water vapor is the same way. It must have some "diffusion velocity" that is proportional to the partial pressure gradient. If the magnitude of the flow velocity out of the leak is greater than the "diffusion velocity," then doesn't it mean that water vapor can't get into the pipe?

Tom

 

[CRG]

tlee123, thinks about the flow velocity profile:

http://www.tpub.com/content/doe/h1012v3/css/h1012v3_40.htm

 

[swainw]

CRG,

Great illustration. The species gradient is inward towards the vapor free stream. The boundry layer at the wall has an effective velocity of zero. This allows molecule migration and then through the diffusion into the flowing stream, getting swept up into. Dry air has a very nice transport phenomenon to "pulling" moisture into its flowing stream. The biggest question will be how much migration happens across the fittings and sealing features that are not exactly perfect.

Bill

Bill Swain
Ultra Electronics Precison Air Systems
Technical Manager-USA
swainw@asme.org

 

[rmw]

As a point of reference, the air in the anecdotal story in the other thread left the paper mill power house at minus 40 degrees F dew point, and the dryer was brand new, just having been installed, and functioning perfectly. Hence the Project Engineers frustration at the complaint ftom the paper machine foreman about the wet air at his location.

I would like to contribute a question to the discussion. Since learning about this type of difussion, I have for years assumed that any steam leak I observed would then, by the same process, be difussing atmospheric air back into the steam piping. Am I right about the same laws of physics working here too? Since all or most all of the air has been removed from the feedwater to the boiler in the deaeration process, the partial pressure of air in the steam is very low, and atmospheric air would want to get in there, wouldn't it? (My question relates to unidentified sources of air in condensers that shouldn't have any air.

One other comment, Thanks CRG for helping me understand why a process that I knew to be true, but which was counterintuitive to me worked. I'll give you a star for that one. SwainW, you get one too for your contribution.

rmw

 

[briand2]

Hospital air conditioning / ventilation is another example. If you ensure that there is sufficient positive pressure in the operating theatre relative to the surrounding dirty areas, you may believe that this also ensures any leaks are from "clean to dirty" so no germs, etc are likely to enter the operating theatre. However, if the vapour pressure of water (humidity) outside the operating theatre is greater than that inside, then water vapour will enter against the air pressure gradient, bringing germs with it!

Brian

 

[rmw]

Unclesyd just posted a link in a chemical engineering thread on a slightly different topic that addresses this issue. See thread124-126159

Look for the paragraph on "Leak Mechansims"

rmw

 

[unclesyd]

Thanks rmw,

I meant to post that on this on this thread for tlee123. Evidently my fingers did the walking without my brain in gear.

 

[quark]

Sometime back, when I started my career with pharmaceutical plants, we had problems with a dehumidified area. The area was maintained at 20⁰C and 20%RH(at higher pressure) and the adjoining room was at 60%RH and same temperature. There was a pass through hatch for container movement between these two areas.

We observed that the RH of the dehumidified area was shooting up to 70% at around 5.00PM. Our initial focus was on the dehumidifier and did what all is required. Secondly, we thought it to be ambient effect. But these things didn't support the effect. One day I was standing infront of the area to observe the rising pattern of RH. I saw a person coming with a pile and scrubber to wet mop the 60%RH area. The RH of the dehumidified are increased and I knew the solution by this time. I spoke to my boss in detail and explained how the partial pressure of water vapor will be high in high RH areas. He never seemed to agree with me, in theory, but there was no change in RH when we closed the hatch.

The same issue was brought up by me two years back in one of the threads but I was ridiculed by an expert(I just don't want to refer that thread here). This prompted me to search for a clue in black and white and I, finally, ended up with Munter's Dehumidification Handbook. The process was clearly explained. The author opines that if the velocity of air exceeds 150fpm, the moisture transfer will be marginal.

But I feel this theory may not support the pipeline issue well. First, though there is a laminar sublayer near pipe wall, even for fully developed turbulent flow, the velocity profile changes as soon as there is a leak. Secondly, the leaking fluid velocities in a compressed fluid system are much higher than 150fpm.

I have seen conditions where the final compressed fluid temperature may go below the DP to which it is dried, particularly near a blocked filter, exhaust and pneumatic controllers(when expansion takes place). This may be one of the reasons for moisture seperation.

Unfortunately, one person, who I believe, can do the calculations of moisture diffusion is out of this discussion. I will just wait.

Regards,

 

[rmw]

Often the leaks in fluid pipelines are at points that are natural flow disruptions, such as flanges, fittings, etc., rather than the classic pinhole in a straight length of pipe.

rmw