[rcastilla]

Hi,

I am working in a simulation with dynamic mesh. But, previously, I wanted to make the stationary simulation. A wall is rotating, and I wonder if I could use the rotatingWallVelocity. I tried with pimpleFoam, and the results are not correct (very high velocities in some cells, near to the rotating wall). Even with potential Foam, it gives velocity zero (no wall velocity?) around the domain, except on the wall.

I checked the rotatingWallVelocity BC source and I found that:


// Remove the component of Up normal to the wall
// just in case it is not exactly circular
const vectorField n(patch().nf());
vectorField::operator=(Up - n*(n & Up));


Does it mean that only the tangencial wall velocity is considered for the fluid motion? Why? Anyway, I will try to remove this constraint, and I see what happens... but I would like to know the reason of that.

Thanks

Robert

[kmooney]

Originally Posted by rcastilla

Hi,

I am working in a simulation with dynamic mesh. But, previously, I wanted to make the stationary simulation. A wall is rotating, and I wonder if I could use the rotatingWallVelocity. I tried with pimpleFoam, and the results are not correct (very high velocities in some cells, near to the rotating wall). Even with potential Foam, it gives velocity zero (no wall velocity?) around the domain, except on the wall.

I checked the rotatingWallVelocity BC source and I found that:


// Remove the component of Up normal to the wall
// just in case it is not exactly circular
const vectorField n(patch().nf());
vectorField:perator=(Up - n*(n & Up));


Does it mean that only the tangencial wall velocity is considered for the fluid motion? Why? Anyway, I will try to remove this constraint, and I see what happens... but I would like to know the reason of that.

Thanks

Robert

I believe that this boundary condition is made for circular boundaries and to act as if it were a no slip no penetration rotating wall, as you described. Because it is a 'wall' no penetration is implied, hence the removal of the face normal component of the velocity. This is a relatively straight forward velocity boundary condition. I doubt that your issues with running it in conjunction with pimpleFoam are related to a bug or something to that nature.

[S03r3n]

Originally Posted by rcastilla

Hi,

I am working in a simulation with dynamic mesh. But, previously, I wanted to make the stationary simulation. A wall is rotating, and I wonder if I could use the rotatingWallVelocity. I tried with pimpleFoam, and the results are not correct (very high velocities in some cells, near to the rotating wall). Even with potential Foam, it gives velocity zero (no wall velocity?) around the domain, except on the wall.

I checked the rotatingWallVelocity BC source and I found that:


// Remove the component of Up normal to the wall
// just in case it is not exactly circular
const vectorField n(patch().nf());
vectorField:perator=(Up - n*(n & Up));


Does it mean that only the tangencial wall velocity is considered for the fluid motion? Why? Anyway, I will try to remove this constraint, and I see what happens... but I would like to know the reason of that.

Thanks

Robert

Did it work to remove the constraint?
Can you share your experience/ results?

Thanks in advance!

[rcastilla]

Originally Posted by S03r3n

Did it work to remove the constraint?
Can you share your experience/ results?

Thanks in advance!

Hi, S03r3n,
It was long time ago, and I don't remember now the details. But I think that we solved the problem finally, and it was published in the Journal of Fluids Engineering, of the ASME:

Castilla, R., Gamez-Montero, P. J., Raush, G., & Codina, E. (2017). Method for fluid flow simulation of a gerotor pump using OpenFOAM. Journal of Fluids Engineering, 139(11), 111101.

It was used to deform the mesh of a gerotor pump.

Regards