Water Bath Fills

Quote:

---

Originally Posted by gschaut

Net sum temperature change (for NEW gas) of zero, and the laws of thermodynamics are preserved.

BUT... in the process of filling your tank from 1200 to 3500psi, the OLD, EXISTING gas that was in the tank (prior to filling)
is compressed as well, and exhibits adiabatic heating.

---

This turns a little light on.

I use an AL80 with a blow nozzle to dry regulator parts. I let the cylinder drain down to about 100psi (or less) before I refill it. Since getting "a good fill" was unimportant, I would just hook it up and fill it quickly. I noticed that going from 100psi to 3000psi, the cylinder didn't warm up very much. But when I filled a customer cylinder (already at 1500psi or so) at that same speed, it got VERY hot.

It makes more sense now.

Quote:

---

Originally Posted by Slüdge

This turns a little light on.

I use an AL80 with a blow nozzle to dry regulator parts. I let the cylinder drain down to about 100psi (or less) before I refill it. Since getting "a good fill" was unimportant, I would just hook it up and fill it quickly. I noticed that going from 100psi to 3000psi, the cylinder didn't warm up very much. But when I filled a customer cylinder (already at 1500psi or so) at that same speed, it got VERY hot.

It makes more sense now.

---

Mumm, not really. It's the mass of the gas, the compressibility and the delta P that determines the heat gain (or loss)

Try this;

Take an al 80 at 2900 (and at ambient temp) and fill it as fast as you can to 3000

Now do the same with an empty al 80.

Which tank gets hotter?

Tobin

Quote:

---

Originally Posted by DA Aquamaster

In an ideal world, when mixing gas for customers, you have the luxury of letting the tanks cool several hours or overnight between O2, helium and the top off gas - but in the real world you need a reasonably accurate mix in a short period of time - an hour or so. I agree that anyone who does any amount of that type of filling and does it well probably understands the effects of a water bath as well as the differences between AL and steel tanks.

But that type of experiential knowledge still does not answer the OP's original request for some hard data on the issue.

---

I think the only useful number would be the difference in dissipation rates between air at a given temp and water @ the same temp. Beyond this there are so many variables that unless you get your fills in a climate controlled laboratory, it would be a case by case basis. Things like water and air temps, humidity, air movement, water movement, the amount of water in the vat, any direct sunlight etc.
Of course, if you get the air cold enough, and get the water warm enough, you could do better filling in the air!:smt102

Quote:

---

Originally Posted by TobinGeorge

Try this

---

That's a little extreme. But under un-scientific conditions, the 1500 - 3000 would always get hotter than the 100 - 3000.

Quote:

---

Originally Posted by Gene Powell

I think the only useful number would be the difference in dissipation rates between air at a given temp and water @ the same temp. Beyond this there are so many variables that unless you get your fills in a climate controlled laboratory, it would be a case by case basis. Things like water and air temps, humidity, air movement, water movement, the amount of water in the vat, any direct sunlight etc.
Of course, if you get the air cold enough, and get the water warm enough, you could do better filling in the air!:smt102

---

Thats' the point of an earlier post with a proposed experimental design - it controls for those variables by keeping them constant rather than trying to identify the effects of each and every one of them. Beyond that I agree to an extent that there are a lot of qualitative factors that come into play.

On the other hand, good fill people do get very good at determining good fudge factors when mixing gas on tight time lines, so while the variable are many, the Mark 1 Mod 0 human brain seems to be pretty good at it, once adequate learning occurs.

Quote:

---

Originally Posted by Slüdge

That's a little extreme. But under un-scientific conditions, the 1500 - 3000 would always get hotter than the 100 - 3000.

---

I would *guess* at high fill rates and large pressure differentials between the source gas and the tank, i.e. near empty al 80, that the tank valve is point of expansion. That produces cooled gas at the tank for the first maybe third to half of the fill, i.e. from 100 psi to ~1500.

When starting with a 1/2 filled tank there is likely much less cooling by expansion at the tank valve.

Tobin

Quote:

---

Originally Posted by TobinGeorge

I would *guess* at high fill rates and large pressure differentials between the source gas and the tank, i.e. near empty al 80, that the tank valve is point of expansion. That produces cooled gas at the tank for the first maybe third to half of the fill, i.e. from 100 psi to ~1500.

When starting with a 1/2 filled tank there is likely much less cooling by expansion at the tank valve.

---

The valve normally is the point of "most" expansion. There is "some" expansion in the plumbing between the cascade bank and the tank being filled, due to pressure loss due to gas friction in the tubing.

The main point here is:

Adiabatic cooling + Adibiatic heating = ZERO (if you are starting and ending at the same pressure).

IF- you take a quantity of gas and reduce it's pressure from "X" to "Y", you get adiabatic cooling.

IF- you take the same quantity of gas and compress it back from "Y" to "X", you get adiabatic heating EQUAL to the adiabatic cooling.

But, tank fill stations are not perfect. As the gas cools as it expands from 3000psi to 1500psi, it picks up heat from the environment.
Then as the gas is RECOMPRESSED back to 3000 psi (as the tank tops off), the adiabatic heating would return the gas back to
it's original temperature BUT.. the gas has been partially warmed by picking up heat fromt the environment (gas tubing between tanks, ect).

So, at the end of the filling process, you end up with a tank that is warmer than the gas that was used to fill it.

Basically, the cascade bank and tank being filled are acting like a refrigerator.

Quote:

---

Originally Posted by gschaut

The main point here is:

Adiabatic cooling + Adibiatic heating = ZERO (if you are starting and ending at the same pressure).

---

So far everyone seems to be focusing on the 1st law of thermodynamics, which in layman's terms says "The best you can do is break even."

The second law of thermodynamics says "You can't even get close".

I think the results of Aquamaster's proposal would be interesting. Who can give it a try? In the spirit of Tahoe, call it the "CDF tank filling benchmark":roll:

Quote:

---

Originally Posted by gschaut

The valve normally is the point of "most" expansion. There is "some" expansion in the plumbing between the cascade bank and the tank being filled, due to pressure loss due to gas friction in the tubing.

---

No it is not, at least when good filling practices are observed.

Virtually all commercial fill stations will have some means to control flow rates *other* than the cylinder valve.

Many have flow controls on the fill whip, and many others will have either pressure regulator and or needle valve well up stream from the cylinder valve.

Pressure drop due to tubing runs from the bank to the reg, whip or needle valve is trivial.
Most fill station use 1/4" od x ~.180" id tubing or larger. Very long runs would be required to induce any significant pressure drop at the modest fill rates seen at the typical LDS.

When filling directly from a compressor, typical of the small home compressor there may not be any flow restrictor, but that's a different set of problems.

Tobin

I would also like to see some final pressure numbers for letting them chill in the water for extra time... say 5 minute increments. There will be some point where it's no longer a huge benefit to keep the tanks in the water.

Maybe we can get a shop to sponsor this test, so we can do a whole bunch of controlled fills. The results could be published on the shops website in exchange.