Thought experiment: Diffusion between tanks

[RainPilot]

Ive always treated gases as very low-density fluids in aviation theory. If I take 2 fluids that are capable of forming a homogenous mixture e.g. milk and water, if i pour one slowly into the other it makes sense that there would be a layer formed. after agitation they would be mixed.

It seems that this type of behaviour is present in a partial pressure blended tank?

 

[huwporter]

Ive always treated gases as very low-density fluids in aviation theory. If I take 2 fluids that are capable of forming a homogenous mixture e.g. milk and water, if i pour one slowly into the other it makes sense that there would be a layer formed. after agitation they would be mixed.

It seems that this type of behaviour is present in a partial pressure blended tank?

Yes, but the thing that is utterly counterintuitive to me, is that filling doesn't create sufficient turbulence to create homogeneity, but shaking the tanks about a bit does.

 

[RainPilot]

Yes, but the thing that is utterly counterintuitive to me, is that filling doesn't create sufficient turbulence to create homogeneity, but shaking the tanks about a bit does.

Maybe because most blenders are adding O2 pretty slowly to avoid ignition?

 

[BenjaminF]

How slowly should they be adding the oxygen? And what sort of buffer is normally used "just in case"?

 

[huwporter]

Maybe because most blenders are adding O2 pretty slowly to avoid ignition?

My most recent example was adding banked nitrox 32 to trimix - not particularly slowly. The analysis was still different after shaking the tanks.

 

[RainPilot]

Fair enough, not a gas blender or flow dynamicist. I imagine various densities at pressure also contribute. I now know to check analysis again when I get home not just the one post-fill, useful tip.

 

[tridacna]

Maybe because most blenders are adding O2 pretty slowly to avoid ignition?

I think that you mean add air slowly. O2 should already be present in the tank.

 

[victorzamora]

Ive always treated gases as very low-density fluids in aviation theory.

...and that's completely valid. In textbooks, fluidics is fluidics. There's a difference between "liquid" and "fluid", though, and that's a trap it seems you've fallen in to. Gases kinda act like liquids, and there's a lot of similarities and they're easier to visualize. There are plenty of areas where the comparison doesn't hold up. The following is an example.

If I take 2 fluids that are capable of forming a homogenous mixture e.g. milk and water, if i pour one slowly into the other it makes sense that there would be a layer formed. after agitation they would be mixed.

It seems that this type of behaviour is present in a partial pressure blended tank?

You've fallen into the trap!!!! The behavior you're describing is due to surface tension. Can you think of a reason why SURFACE tension isn't present in gasses?

Gas has no surface.

 

[victorzamora]

I think that you mean add air slowly. O2 should already be present in the tank.

@RainPilot ....to expand on this, shops that use Partial Pressure fill methods typically pump O2 in first. This allows them to transfer O2 at lower pressures and allows them to get more out of the bottoms of their O2 sources. That gets pumped slowly. Filling it up after that can be done at any speed the shop is capable of. One note, though, there are a lot of gas-blending methods and I think only Partial Pressure fills are subject to this kind of mixing. Almost every Nitrox fill I've ever gotten has been out of banked gas, though, where that's not really an issue. Continuous blending is another fill method that doesn't have this stratification issue.

 

[tridacna]

Is there any way to predict or model the mixing behaviour of two gases?