Thoughts on creating a flow inside a channel

Hello everyone,

I have two solid walls and a liquid in-between the walls. I want to make the liquid flow in to a one direction, something similar to a Poiseuille flow. Yeah, the LAMMPS example script “in.flow.pois.lmp” based on the “fix addforce” works fine. But, I thought what if I could move the two walls (the top and bottom) to a single direction instead of the liquid. So, relative to the walls, the liquid should flow to the opposite direction. However, when I simulate it using the “fix move” on the walls, the whole system (walls+liquid) goes to a single direction (to the direction I specified in the fix move). I’m confused here. Can anyone comment on this?

Thank you.
Chamara

Hi,

If the walls are moving in one direction, and the liquid is in contact with the walls, then the liquid will end up moving in the same direction as the walls. I see nothing strange here. What other possible scenario did you expect here? Why would the liquid flows in the other direction?

Simon

Hello Simon,

Thanks for the reply. Yeah, you’re right, liquid should start moving with the walls due to the inter-atomic interactions. But, shouldn’t be there a velocity profile as shown Fig. 1 (a) within the liquid? Therefor, relative to the wall, the liquid should flow on the other direction as shown in Fig. 1 (b).

Chamara

No. Those velocity profiles are not realistic (note: Fig 1a may describe a situation that is realistic just after the walls start moving, i.e. before the fluid reaches its stationary velocity profile).

Here, when the walls are being moved at constant velocity, the situation that would be the most favorable energetically is when the water forms a plug flow, ie when the entire water layer moves at the same velocity as the walls (because gradients of velocity as you draw means dissipation of energy).

In nature a Poiseuille flow arises from a pressure gradient. The wall-fluid slip is a consequence of the Poiseuille flow, not its cause. So it’s quite nice to see that your LAMMPS simulations obey physics in this case!

Dear Simon,

Thank you very much for your insighful explanation. Now I can see where I went wrong. In a Poiseuille flow, the chanells are fixed (we are holding the chenells) while the liquid flows inside. But in my case, no one is holding anything, so the fluid accelerates to the same velosity as the wall as you discribe.

Thanks again. Cheers.
Chamara

Hello srtee,

Thanks for your informative reply. Now I can see why my expectations are off.

Cheers.
Chamara