[mostanad]

Hi everyone,


I hope you are all well. I have a CFD question about my simulation. I have a suspension flow with unbounded condition. This condition enforces the bulk velocity (the volumetric particle and fluid velocity) to be zero. So because of this, the particles in the suspension settling down while the fluid goes up. I have done this with periodic boundary conditions entire the domain.




On the other hand, I have tested my domain once again with only wall BC in one direction. The other directions are still periodic with unbounded condition. So when I look at my results:





It seems that the mean particles velocity, because of the wall shear stress, has a lower value, which is reasonable.

But I wanna minimize this wall effect through moving it upward. I mean having moving wall but with a result close to periodic boundary condition. What do you think about the velocity value for the wall? What criterion should be considered for that?

[FMDenaro]

Originally Posted by mostanad

Hi everyone,


I hope you are all well. I have a CFD question about my simulation. I have a suspension flow with unbounded condition. This condition enforces the bulk velocity (the volumetric particle and fluid velocity) to be zero. So because of this, the particles in the suspension settling down while the fluid goes up. I have done this with periodic boundary conditions entire the domain.




On the other hand, I have tested my domain once again with only wall BC in one direction. The other directions are still periodic with unbounded condition. So when I look at my results:





It seems that the mean particles velocity, because of the wall shear stress, has a lower value, which is reasonable.

But I wanna minimize this wall effect through moving it upward. I mean having moving wall but with a result close to periodic boundary condition. What do you think about the velocity value for the wall? What criterion should be considered for that?

Why don’t you simply set the wall stress to be zero?

[mostanad]

Originally Posted by FMDenaro

Why don’t you simply set the wall stress to be zero?

Dear Filippo,
Thank you for your reply. You mean free-slip BC?
So my problem is eith its definition. Is this really a wall? I can't comprehend it physically.

[FMDenaro]

Do you think that two sliding walls are more realistic? I think that your problem is the wall confinement, you should check first the streamlines for the fully periodic case, then you have to fix the position of the wall in a location where the streamlines are vertical (or as much as possible). Then, you can also set a free-slip condition.

[mostanad]

Originally Posted by FMDenaro

Do you think that two sliding walls are more realistic? I think that your problem is the wall confinement, you should check first the streamlines for the fully periodic case, then you have to fix the position of the wall in a location where the streamlines are vertical (or as much as possible). Then, you can also set a free-slip condition.

Ok. But I wanna have the wall again in the same place. But my non physical understanding of free-slip condition, then we will have a wall that its varying velocity value at each nodes. Since every node at the wall want to have the velocity of its neighbouring inner node. So can we regard this as a wall?

[mostanad]

Originally Posted by FMDenaro

Do you think that two sliding walls are more realistic? I think that your problem is the wall confinement, you should check first the streamlines for the fully periodic case, then you have to fix the position of the wall in a location where the streamlines are vertical (or as much as possible). Then, you can also set a free-slip condition.

No opinion in this regard, dear Filippo?

[FMDenaro]

Originally Posted by mostanad

Ok. But I wanna have the wall again in the same place. But my non physical understanding of free-slip condition, then we will have a wall that its varying velocity value at each nodes. Since every node at the wall want to have the velocity of its neighbouring inner node. So can we regard this as a wall?

I don't understand, the free-slip means that your wall is at the rest, the fluid will have a non-zero tangential component along it, that should mimic your fully periodic case.

[mostanad]

Originally Posted by FMDenaro

I don't understand, the free-slip means that your wall is at the rest, the fluid will have a non-zero tangential component along it, that should mimic your fully periodic case.

OK. Now just imagine I wanna lower the wall effect in the suspension direction. This can be done using the free-slip BC. However, I wanna apply a shear on the suspension on the cross-direction as well. I mean no wall effect on z direction (gravitational direction), but applying a shear in the y direction. So then you think can it be done using a free-slip BC along with a body force term in y direction?
Sincerely,
Mohammad

[mostanad]

Originally Posted by mostanad

OK. Now just imagine I wanna lower the wall effect in the suspension direction. This can be done using the free-slip BC. However, I wanna apply a shear on the suspension on the cross-direction as well. I mean no wall effect on z direction (gravitational direction), but applying a shear in the y direction. So then you think can it be done using a free-slip BC along with a body force term in y direction?
Sincerely,
Mohammad

I will be happy if you, Filippo, can help me with this idea.
Cheers,
Mohammad

[FMDenaro]

Could you show your resolved velocity field in the fully periodic case?

[mostanad]

Originally Posted by FMDenaro

Could you show your resolved velocity field in the fully periodic case?

Sure,
First I should mention that settling happens in z direction.

This is the velocity magnitude field for mid-plane in the tri-Periodic case:





I have also plotted the velocity vector over this field. The regions with shaded horizontal vectors are the particles region.


For wall in x-direction, however, it changes to the following field:





Besides the particles which are free to cross the periodic x-direction, I need to minimize the stress force imposed over the field along the suspension direction, z. So, this can reduce the confinement effects caused by the stationary wall.



As I showed you in the initial post, the wall reduces the particles average settling velocity, which because of the unbounded domain, it also reduces the upward fluid velocity. I need to have a situation in which both the settling particles velocity and upward fluid velocity are less affected by the wall.

What do you think about this?

[mostanad]

Originally Posted by FMDenaro

Could you show your resolved velocity field in the fully periodic case?

Dear Filippo,
Do you have any opinion about this? Your help is much appreciated.
Mohammad