liquid level sensing - potential leak

[skacur]

we are experiencing what appears to be a slow air leak in the following setup:

 

low pressure sensor connected to a tygon tube via 1/8" hose barb; the tygon tube is submerged in a liquid; the hydrostatic pressure of the liquid compresses the air in the tube; the sensor measures the air pressure; the linear relationship between hydrostatic pressure and liquid level enables liquid level determination

 

problem: over the course of several weeks we've noticed a slow, steady drop in pressure while a constant liquid level is maintained. of note, the small "slug" of water at the bottom of the tube climbs slowly and steadily indicating an air leak.

 

two questions:

1 - if this is an air leak, is it through the sensor, the hose barb seam or the tubing itself (permeation)? we have tested several variations of this with no conclusive answer

 

2 - if this is not an air leak, what is causing this phenomenon?

 

thank you in advance

[Mr Skeptic]

Well, when fixing bicycle tires, if I had trouble finding a leak I would put the tube underwater and look for where the bubbles come out. Can you pressurize and submerge the system? If your system is waterproof you could put in water or some other liquid that is safe for it, and check for leaks. You could also put in a gas that you can detect (eg by smell) and then find the leak.

 

Would any of these suggestions work?

 

Oh, and there hasn't been any steady decrease in atmospheric pressure has there? Nor any cooling of the air in your system?

[Externet]

Can be both; 1- and 2-

2- being probably air dissolving into the liquid. As a fish breathes the atmosphere dissolved in water.

A better way would be a submerged deformable sealed 'bladder' instead of a tube that allows an interactive surface between air and liquid.

Edited November 21, 2009 by Externet

[josh richards]

Could a change in the atmospheric pressure outside of your closed system be a suspect?

[skacur]

Thank you for the responses.

 

Regarding atmospheric pressure, there are fluctuations, but not a steady decrease (or increase). Also, we are using a differential pressure sensor which is referencing the atmospheric pressure, so the sensor output should not be affected by these atmospheric fluctuations.

 

Likewise with temperature, there are small fluctuations (68°F to 75°F) but no steady increase or decrease.

 

We've tried to pressurize the setup in a bath of water and have found no visible bubbles. We've done this to the 3 psi limit of the sensor. In operation, the sensor will see no more air pressure than 12" of water.

 

We have actually tried a latex diaphragm/bladder, but have found that the response and accuracy are greatly diminished. Maybe there are better materials/geometries.

 

Externet: The air dissolving in water is an interesting point. Do have any formulas to predict the extent of this?

 

Thanks again.

[Mr Skeptic]

Externet: The air dissolving in water is an interesting point. Do have any formulas to predict the extent of this?

 

Hm, I hadn't realized that one end was open to the water. The amount of air you lose due to dissolving in water will depend on the amount of air already dissolved in the water, and the temperature. Colder water dissolves more gas (if this seems counter-intuitive it is because the gas molecules need to slow down to enter a liquid). Other things dissolved in the water could also affect it. If you water is warmed then cooled it will definitely be able to dissolve some air.

 

Henry's Law tells you how much gas can dissolve. However since there is little surface area the rate of dissolution will be rather slow.

 

Regardless of the source of the problem, if you are losing air you can replace it by removing the tube from the liquid and replacing it. The amount of air in the system will then be the volume of the system times the density of air at the current atmospheric pressure.

[Externet]

Hi.

Formulas and calculations for the slow tube air dissolving into the liquid could be better exposed by someone with deeper scientific background as many are qualified in this forums, considering the temperatures and atmospheric pressure variations, and area of the liquid-to-air surface inside the tube.

I will stay in the qualitative corner.

 

Occurs to me that if the liquid is periodically (daily?) emptied and immediately returned to the vessel, the sensing tube will renew its air level to confirm or discard if my 'dissolution' suspicion is the reason for the gradual discalibration.

 

Another way would be to briefly remove and return to submerge the sensing tube to 'reset' its reading. There is another way, by (daily?) purging from an upper 'T'

 

As a domestic washing machine level control that uses the same principle, every time it is emptied, the tube fully drains. Every time it is filled, has a reset point to start sensing its water level by sensing the air pressure in it.

 

Miguel

[skacur]

Unfortunately, we can't rely on our customers removing/resubmerging the tube on a regular basis. Likewise with the purge method.

 

I took a quick look at Henry's law. It looks like solubility is dependent on air pressure and liquid temperature. Does anybody know how the surface area of the air/water interface is accounted for?

[CaptainPanic]

Doublechecking if I understood the question

From what I understand: you have a liquid pool. A pipe of some kind that is closed off on one end with a pressure meter, and is open at the other end. And you stick that pipe into the pool.

 

The deeper, the pipe (or the higher the liquid level) the higher the pressure of the gas in the pipe.

 

This means you always have a higher pressure in the pipe than the outside air...

 

Answer

If the gas (air) is in direct contact with the liquid phase, air will dissolve.

 

Henry's law will tell you the equilibrium state: if you have a certain gas pressure (approx. 1 bar), it will tell you the concentration of the gas in the liquid.Note that the pressure of the air in contact with the liquid is different on 2 places: inside the pipe and the outside air.

 

The liquid will be saturated with air, meaning it is at equilibrium... it's at equilibrium with the outside air!

 

But, the pressure on the side of the pipe is higher than on the side of the outside air! So, the concentration of the air dissolved in the liquid on the side of the pipe is higher than on the side of the outside air!

 

So, we have a single liquid phase which has more air dissolved on one end than on the other. What happens next is that air will slowly diffuse through the liquid to where the concentration is lower... This in turn enables the dissolution of more air from the side of the pipe.

 

After the air diffused all the way through the liquid phase (possibly there is even normal mixing of the liquid, which will speed things up?), the surface in contact with the outside air has a higher concentration in air than the equilibrium value... and air will evaporate.

And that's how you lose your air.

 

Remember:

A liquid is not air tight if you give the gas enough time to dissolve, diffuse, and evaporate again.

 

And the bigger the surface area (the contact between the air in the pipe and the liquid), the more air you can dissolve per time.

 

There isn't much you can do about this I'm afraid. You must physically prevent the air from dissolving. Use a piston or a plug, or something like that.

 

p.s. This is a typical chemical engineering problem. If this would have been put in engineering, insane_alien or I would have picked it up faster.

Edited November 25, 2009 by CaptainPanic
I pushed the post-button too soon. Wasn't finished yet :D

[skacur]

Captain Panic: Thank you for your response. Your explanation makes perfect sense. I am struggling with one caveat however. We've submerged an inverted glass test tube to create a hermertically sealed chamber. This setup has shown no evidence of air escaping over several weeks (i.e. the water slug at the bottom is not climbing). I took this as evidence that we are experienceing a leak somewhere in our plastic setups. Let me know what you think.

[CaptainPanic]

Parameters that are of importance are:

 

- The depth of the water level in the tube and the pipe should be the same - to generate the same pressure.

- The length of the test tube / pipe that contains water (the water level rises inside the pipe) must be the same. The air will slowly diffuse through the nearly completely stagnant water in the pipe, and then once it's in the "bulk" of the tank, the mixing will transport the air away quickly. So, the longer the pipe that contains water, the slower the air will diffuse out.

 

That is also a solution to the problem I described in my previous post.

 

Other than that, I have no answers. It might still be a very small normal leak... a little scratch on your hose barb can cause a tiny leak. If something as little as 1 ml of air per day is a problem, I have no idea how to test it for leaks.

[J.C.MacSwell]

Any chance the tygon could be absorbing water and expanding?

 

The glass of course would not do this and the geometry would stay consistent.

[skacur]

J.C. MacSwell: I looked into the water absorption of Tygon and it is about 0.2% after 24 hours. I haven't found what the saturation amount would be but this looks to typically be about 2X the 24 hour amount for most plastics.

 

I think if the tubing were expanding, increasing the volume of the tube, this would explain the water slug climbing from the bottom since there would be the same amount of air in a larger tube. However, if the tube were to grow in length, the apparent depth would be greater and the pressure reading would increase. This is not the case, the pressure reading drops consistently over time.

[skacur]

Does anyone happen to know if there are any coatings that can be used to improve the gas permeability of Tygon? I've been looking into CVD(chemical vapor deposition), PVD(physical vapor deposition) and related technologies. There are plastic beverage bottle manufacturers that are using coatings, but I haven't found any contract manufacturers that do this work. Any ideas?