SpudFiles

At the office I have a few air tanks that combine for a volume of roughly 1200 cubic feet. They are normally used for the storage of 3000 psi air (Note: When I say they're 1200 cubic feet, I don't mean that they hold 1200 scf of air; I mean that they have a combined volume of 1200 cubic feet.).

In any event, despite the use of air dryers, we get condensation in the tanks. This is bad (mmmkay?).

What I'd like to do: Periodically vent the tanks and then pull a vacuum on them to force any condensation in the tanks to boil off thereby drying the tanks out.

Early in the process this should be easy, but as the vacuum hardens (read: pressure reduces to < 0.5 psia) things could get dicey. More to the point, it is my understanding that pulling high concentrations of water vapor through a vacuum pump can destroy a vacuum pump. CAN destroy. Not WILL destroy. Again, it's my understanding that the difference between "can" and "will" in those two sentences is primarily depending upon the type/quality/etc. of vacuum pump.

I confess to knowing virtually nothing about vacuum pumps.

Thus, with all that said I am asking: Does anybody here know enough about vacuum pumps to recommend one for my situation/application?

We have liquid ring pumps that don't care about water at all, actually they welcome it

http://en.wikipedia.org/wiki/Liquid_ring_pump



Downside is it gets very hot so normally you need to recirculate the water, possibly with a heat exchanger involved - so there is some plumbing to do, unless you get a sealed one but I have no experience with these so can't comment.

In our case we stopped using them because during a distillation the vapours that escape the condenser end up mixing with the water in the pump, which then means the water has to be treated as contaminated waste, which is a pain in the posterior.

You could always use a "normal" oil pump and add a water trap.

Having some exposure to your requirement, here are the issues.

1 Most air gets pulled off quickly.
2 Remaining water boils or phase changes to water vapor.
3 After the pump, the phase change reverses leaving the pump contaminated with water.
4 Flow rate at low pressure is low.

As a result pulling a vacuum is not the ideal solution to drying tanks. The process takes a very long time.

The solution is not difficult.

1 Raise the vapor pressure of the tank. In short, bake it. Elevated temps from 175-250F work wonders in increasing the vapor pressure.
2 remove the vapor and reduce the water vapor pressure to encourage evaporation. In short, purge cycles with dry anything in a heated tank purges out water vapor. A vacuum pump can be used for the cycle process.
3 To protect the vacuum pump, if you need to pull a hard vacuum (no purge gas leaving only water vapor) a cold trap is recommended. Place a length of the vacuum line in a bath of alcohol and dry ice to condense the water vapor and phase change it to a solid before it reaches the pump.
4 Use at least 3 pump purge cycles to dry a tank at elevated temperature. This is called bake out in high vacuum equipment.

Google high vacuum bakeout

Jack,

As I understand it, the water-based liquid ring pumps don't pull much of a vacuum. True/false?


Tech,

Look at my volumes. We're talking about 8 tanks that have exterior dimensions on the order of 4' diameter x 30' long. To add insult to injury, they're outdoors. Heating them is not a realistic option.

The cold trap sounds interesting but may be a PITA to get through the environmental office.

To add insult to injury, they're outdoors. Heating them is not a realistic option.

Painting them flat black could help if you're in the desert... of course that might not be an option if pressure fluctuations are not tolerable during 'normal' use.

You could arrange the plumbing so that the air being sucked out is bubbled up through a bath of water-mix cutting fluid(or similar), which would inhibit corrosion if it got into the pump. I don't know how the vapor pressure of cutting fluid compares to water, but I would imagine varies somewhat by brand and intended application.

@D_Hall

BPA uses them to dry new transformers in substations prior to adding the dry Nitrogen to keep moisture out of the transformer oil. Oil does draw some moisture. Initial moisture in oil in the presence of high voltage can cause failure. EPA approval to remove moisture from your pump forline should not be any more difficult to get approval than simply pumping the moisture to the atmosphere. In one you make ice to keep it from the pump.

Another use for them is in the food freeze drying industry.

Fnord,

Yes, pressure fluctuations are bad. During the summer it would be easy for the tanks to see a 100 F temperature fluctuation from night to day. That would result in a (roughly) 20% pressure increase. Start at 2850 psi at sunrise and find yourself at 3400 by mid afternoon. The safety valves blow at 3100.... And when they blow... Well, you're down for a couple days.

As for the bubbling through cutting fluid... Interesting. I'll have to give some thought.


Tech,

When I referred to our environmental office I was primarily concerned with red tape in general and "what do you do with the alcohol when you're done with it" concerns. Good news I learned today, however, is that apparently we have a dry ice plant on site (albeit a few miles away). Good to know!

D_Hall wrote:As I understand it, the water-based liquid ring pumps don't pull much of a vacuum. True/false?

We manage -1 bar in a 2000 litre reactor in a matter of minutes with one pump runnung (don't remember the specs, I can check if you want), and I can tell you the operators were quite irked when we switched to oil pumps, along the lines of "screw the environment, we want our powerful vacuum back!"

Your volumes are significantly bigger but we have these pumps running 16 hours at a time, they don't mind as long as you keep them cool.

When I referred to our environmental office I was primarily concerned with red tape in general and "what do you do with the alcohol when you're done with it" concerns

We use condensers that work pretty much like simple laboratory condensers - coolant going through a coiled tube.



We have two options for coolant, either mains water, or a mixture of water and monoethylene glycol (abou 15-20%) run through a chiller. Obviously the latter is more effective, but more complex to set up and service.

Again though with a liquid ring pump, a trap like this is redundant if all you're doing is getting rid of moisture because feeding it directly into the pump causes no issues whatsoever.

I didn't explain pump purge theory well.

If you have a pump that can pump to .1 atm, for example, then running it overnight on a damp tank will pump down the tank to 0.1 atm. The pressure of any mixture is the sum of all partial pressures present. Air would be about 20% Oxygen and 75% Nitrogen + water vapor and other gases such as CO2 at the beginning of pumping.

Water in the tank would change to vapor as the pump continued to run. Soon the air is blown out of the tank by water vapor.

Now the tank has almost all water vapor at 0.1 atm.

If we fill the tank to 1atm of dry nitrogen, the Nitrogen will mix with the 0.1 atm of water vapor so the tank contains 0.1 atm water and 0.9 atm of Nitrogen. Pump the tank again to 0.1 atm pressure. This removes 90% of the remaining water vapor pressure to bring the tank to 0.01 atm of water vapor pressure. Repeat 2 times more.

At that low vapor pressure, any remaining moisture is evaporated and purged.

Pump purge can do much better than a long pump cycle alone.

A trap will work but to get it to actually pull water out of a low pressure stream you'll need a seriously low temperature. In a chem lab the trap would be immersed in dry ice acetone (or ethanol) and be at about -70C. That's really the only way to keep moisture (and other volatiles) out of the pump. With this setup you basically distill the water from the source system into the trap and you can get virtually 100% of the water out of the source system given enough time.

A high vac pump (bank off pressure down around 0.002mmHg, 0.002mBar) will pull water out of any system regardless of temperature as long as the temperature is reasonable (e.g. > -10C). Indeed in a lab a "lyophilizer" uses hard vacuum to pull large quantities of water out of a sample at 0C where the water is frozen. The refrigeration system on lab lyophilizers cool the condenser to about -45C and a hard vacuum pump used.

One possibility is to use a vac pump that runs at fairly high oil temperature. The hot oil will tend to auto-purge itself of water. Direct drive pumps often have oil temperatures during continuous running of perhaps 150F. That helps the oil purge itself of water. Belt drive pumps usually run cooler (and generally last longer between oil changes) and don't auto-purge as well. A direct drive pump like http://www.labconco.com/_Scripts/Editc20.asp?CatID=25 might work, though ~50 gallons/minute for free air might not be enough for your huge tanks.

Tech's idea of using a nitrogen (or argon) stream is a good one. You can use a high volume pump that doesn't pull all that "hard" a vacuum. Compressed nitrogen is pretty cheap (and generally very dry) but given the huge size of your tanks might not be practical.

The cheapest approach might be to just wait for a hot day and use a high volume low pressure compressed air source to just blow lots and lots of air through the tanks. When the bottom of the tanks no longer feel cool to the touch you've probably removed the majority of the moisture.

Tech,

I got what you meant with your pump purge theory. We do a similar process (albeit reversed) when we bring our oxygen systems online. We obviously start with air. We then fill to 300 psi with boiled off LOX. We dump the tanks. We repeat a couple times, then we inflate to 3000 psi.

With such a system (albeit using negative pressures) I can see how all vapor is removed easily. I'm just concerned about liquid phase water and it's rate of evaporation when ambient temperatures are sub 40 F.

As for the desire for a single long cycle.... I just figure if that'll work it'd be easiest as I can simply start the pump on a Friday evening before I leave and turn it off on Monday morning when I return. It's obviously not as efficient, but it sure is easy.


Jimmy,

Thanks for the tip regarding hot oil purging the water. That totally makes sense.



All,

Along the lines of the soft vacuum and purge concept. I'm wondering... If I'm not going for a hard(ish) vacuum, might a simple eductor be suitable? I've got a portable 140 CFM air compressor that I could use to drive one quite easily... If there's a point in it.

despite the use of air dryers, we get condensation in the tanks.

just out of curiosty...
wouldn't it be a better idea to circulate compressed air through air dryes when it is at high pressure and condensation occurs at a fast rate?

Sure that would mean you need a small compressor that takes air from one side or the tank, directs it to an air drier and returns it to the tank - ie. it would have to run continously but you wouldn't need anything more than just a few psi to pump air through the dryer

edit
ohh and just to be sure - are the dryers drying high pressure air or are they located between for example the first and the second stage?

POLAND_SPUD wrote:just out of curiosty...
wouldn't it be a better idea to circulate compressed air through air dryes when it is at high pressure and condensation occurs at a fast rate?

Sure that would mean you need a small compressor that takes air from one side or the tank, directs it to an air drier and returns it to the tank - ie. it would have to run continously but you wouldn't need anything more than just a few psi to pump air through the dryer

Depends on the dryer technology in use and the ambient conditions. Desicant-based dryers would certainly work in that manner but they tend to be high maintenance and very fragile. We're working in a high vibration environment (as in, the very ground shakes) so we stick with refrigeration-based systems. They work great to about 40 F, but below 40 F... Condensation.

ohh and just to be sure - are the dryers drying high pressure air or are they located between for example the first and the second stage?

Our dryers operate at 3000 psi. They're between the fifth stage and the tank.