Gas mixing physics

[oxyhacker]

Sound typical of the hyperbole that gets said when one has a product to sell. If he really ever saw a 30% error he ought take a basic mix course or read a good book or two before trying to write about it himself.

How long it takes a mix to homogenize is, interestingly enough, still a somewhat controversial subject. Nitrox seems to mix pretty quickly, He a little slower but I've never noticed much change after, say, 2-4 hours, but others swear they do for as long as 12 so it makes sense to wait as long as possible. Don't know if the differences are due to hardware, technique or just imagination, but this is not something you have to take on faith and I always suggest that new mixers sample their mixes at intervals over 24 hours, until they got enough of a feel for it to reach their own conclusions. In the old days you used to hear all sorts of strange stories, like about how changing the position of the tank while analyzing would change the reading, as the He floated up or down, and how the He would leak out selectively causing the FHe to drop over time. You don't hear these as much anymore, but there's still far from agreement on how long to wait

Oh, it's odd that while in the quote he seems to be pushing the analyzer as a solution to the real/ideal gas conflict and the problem of slow homogenization it will in reality do neither - you still have to compensate for compressibility when you do your premix calculations, long before the mix can be analyzed, and an He analyzer will do nothing to make the gases homogenize faster! It will help detirmine when the mix has finally stabilized, but so will an O2 analyzer for a fraction of the price.

cyklon_300 once bubbled...
Posters on the Nitrox Equip thread were discussing the rate at which high pressure gas mixtures reach a state of homogeneity. Some (including me) were under the impresssion that this occurred in a very short time period (approaching instantaneously) in a scuba cylinder.

Others stated that they allowed periods of time and/or physically agitated the cylinders to promote reaching equilibirium.

I just found the following statement concerning trimix which I thought was quite interesting:

"It is now evident that helium based mixes are extremely vulnerable to real gas mix laws and can be more than 30% in error and/or can take more than 12 hours to reach a homogenous state. This robust, reliable unit quickly determines exactly what gas mixing has occurred and is an inexpensive and mandatory component for every mixed gas application. I will never again use a mix, or allow a mix to be used by a student or customer without first testing it with the helium analyzer. To do so otherwise would be irresponsible and reckless.” - Bill Nadeau

Bill Nadeau is the Proprietor of Deep Fathoms in Nanaimo, British Columbia. Bill has authored many texts on Scuba Diving including the IANTD Gas Blending Manual and IANTD Mixed Gas Blending Training Slides. He is also a member of IANTD Board of Advisors and an IANTD Trimix and Trimix Blending Instructor Trainer Trainer (ITT). Some of his many distinguished achievements include:
British Columbia Safety Council on Diving Advisory Board Member
First Mixed Gas Cave Dive in Canada - 1998
Deepest Cave Dive in Canada - 262ffw - 1998

The statement of "12 hours to reach a homogenous state" is what caught my eye.

Any physicists or gas blenders care to comment?

 

[donacheson]

cyklon_300 once bubbled...

The statement of "12 hours to reach a homogenous state" is what caught my eye.

Any physicists or gas blenders care to comment?

Strikes me as very conservative. Speaking as a physicist, the mixing rate depends primarily upon conditions affecting turbulent mixing, not molecular diffusion. However, those conditions are not easily controlled. I'd also expect rolling the tanks to be inconsequential.

In any case, measuring the mixture would, IMHO, be mandatory.

 

[pescador775]

If diffusion is desired perhaps some enterprising person can come up with a tank banger, something other than a hammer, that is. There is a barnacle remover which emits sonic pulses......on the other hand, one could carry tanks to a rock concert.

 

[scorpionfish]

Uncle Pug once bubbled...
What exactly are the "real gas mix laws" refered to in the first post?

Hi, Uncle Pug!

In the "real world" there is/are the Ideal Gas Law(s), the most fundamental of which is that PV = nRT. That simply states that there is a relationship between the pressure, volume, quantity and temperature of an "ideal gas" The "R" is a constant that can be thought of as a very sophisticated "fudge factor". If all gases were ideal, they would obey this law.

Of course, not all gases are ideal, so chemists, physicists and engineers have come up with some even more clever fudge factors called by the unlikely name of "fugacity"--and it is not an R-Rated term. Now, all this has nothing to do with the rate of mixing, just the ratio of mixing!

When it comes to the rate of mixing there is one variable that is paramount: turbulence. If the gases are delivered to a vessel simultaneously at a high rate, there will be sufficient turbulence to effect good mixing. There would be a problem if the gases were delivered separately and slowly from a single point. Now one would have to depend on diffusion to mix the gases--and diffusion is a slow molecular process.

What speeds it up? Temperature and physical mixing. Would rolling a tank help? NO! That really only works with solids or liquids that do not fill the whole space of a vessel. To mix gases in a closed vessel with no internal agitation, you have to rely on diffusion--and only higher temps aid that process.

So. the answer is, since most places feed gases sequentially using the partial pressure method, there could be segregation. However, usually they use a high enough rate to cause turbulence and the proper blending is effected. How can you be sure? Analyze it! Is it a problem? Probably not. However, don't you always do an analysis on your gas mixes? if you are anal-compulsive like me, you do!

Scorpionfish

 

[Uncle Pug]

scorpionfish once bubbled...
Hi, Uncle Pug!
don't you always do an analysis on your gas mixes?

Hi, Scorpionfish! Nope. Not once it's in the jug.

But then I continuous blend and analyze every so often during the fills.

The turbulence in the blender as well as compressor seems to take care of mixing things up thoroughly. Analysis later shows no change.

Occassionally I analyze the bank to make sure that there hasn't been any mix creep over multiple fills.

 

[scorpionfish]

Uncle Pug once bubbled...

Hi, Scorpionfish! Nope. Not once it's in the jug.

But then I continuous blend and analyze every so often during the fills.

The turbulence in the blender as well as compressor seems to take care of mixing things up thoroughly. Analysis later shows no change.

Hi, Uncle Pug!

Well, you may not be as compulsive as I (few are), but you are a close second with your carefully thought out fill technique! Pug's Jugs for all you Mugs--didn't Edgar G. Robinson say something like that--or maybe it was one of the Bowery Boys???

Scorpionfish

 

[DameDykker]

If you'll try to calculate the average time for the gas mixture to be homogenous the flow pattern is important. At very low rate (Reynold's number low) you'll have laminar flow and have to rely on diffusion to get the mix, as perviously mentioned. At higher rates (high Reynolds number) you'll have turbulent flow and comparetive high mixing speeds. Re=v*D*eta/rho (Reynolds number =velocity*diameter*viscosity/density). With non-ideal gases the viscositis and densities will vary with pressure and mixture so to solve that you'll get (together with the real gas law) a bunch of connected differntation equations to solve which should probaply be done numerical. So maybe it's eacyer just to measure the mixture over time - i.e. the engineer approach!

 

[Stone]

DameDykker once bubbled...
So maybe it's easier just to measure the mixture over time - i.e. the engineer approach!

DameDykker,

I am appalled that you would call this the "engineer approach". As an aerospace engineer, I assure you that I would use all due diligence in calculating every viscosity and density of the subject gasses before applying the appropriate coefficients. Only after I was completely satisfied would I ask a "technician" to measure the mixture.

P.S. Now the technicians on the board can make fun of someone else.

 

[scorpionfish]

Stone once bubbled...

DameDykker,

I am appalled that you would call this the "engineer approach". As an aerospace engineer, I assure you that I would use all due diligence in calculating every viscosity and density of the subject gasses before applying the appropriate coefficients. Only after I was completely satisfied would I ask a "technician" to measure the mixture.

P.S. Now the technicians on the board can make fun of someone else.

Yes! As a chemical engineer, I, too, was shaken by DD's answer. Everyone knows that the good engineer does exactly as Stone describes: calculate from first principles, recalculate using appropriate coefficients (known to some as "fudge factors"), and, somewhat finally, optimize. Only after all those steps does one, then do the final step: turn it over to a technician for measurement--and then, perhaps, create some new fudge--er--coefficients! Well, I am on the boundary layer on this one!

Scorpionfish

 

[DameDykker]

Woke you up properly, didn't I

Being both and engineer (chemical) and a technician I would probably do both. But not having eacy access too all the viscosity and desity figures and probably not wanting a general solution but a specific one - well then i'll just go ahead and measure it. :bonk:

You can start throwing cheap and heavy object in my genneral direction now!