[wrcase]

Dear FOAMers:
I am having difficulties applying the fvOptions file (new to OpenFOAM 2.2) to do interFoam calculations in a rotating framework.

To get a better understanding, I created a minimum working example (MWE), where I defined a graduated cylinder (100 mm tall x 20 mm radius) using axisymmetry with the centerline following the y-axis. The cylinder is filled 50 mm with water and open to atmosphere on top as shown in the first image (plot of alpha1 variable). I then rotate the cylinder about the axis of symmetry and would expect the water level to rise on the side, but lower in the middle. However, I seem to get no response. In the second image, I should see a rise in pressure at the right side wall. Also attached is the case directory.

alpha1.gifpressure.gifcylinder.tar.gz


Questions:
Q1. Am I defining the cellzones correctly in blockMeshDict (first in blocks statement and 2nd in boundary statement)? Example:

blocks
(
    hex (   0   2   4   1   0   3   5   1) rotor (  10  50   1) simpleGrading (1 1 1)
);
boundary
(
    wall
    {
        type wall rotor;
        faces
        (
            ( 0 2 3 0 )  // Bottom
            ( 2 4 5 3 )  // Right
        );
    }
    atmosphere
    {
        type patch rotor;
        faces
        (
            ( 1 1 5 4 ) // Top
        );
    }
...

Q2. Am I correctly assuming the fvOptions is the way to add Coriolis and Centrifugal force terms to my momentum equation. Keep in mind that in my actual problem both play a role. Am I correctly describing this?

MRF1
{
    type            MRFSource;
    active          true;
    selectionMode   cellZone;
    cellZone        rotor;

    MRFSourceCoeffs
    {
        nonRotatingPatches ();
        origin      (0 0 0);
        axis        (0 1 0);
        omega       constant 50;
    }
}

Q3. Is the variable omega the rotational speed in rad/s? If so, then I have about 5 g's at the right wall, so I should definitely see something.

Thank you for your support,
Bill

[vonboett]

Hi Bill,

If you are in a hurry, a work around would be using interDymFoam with rotating the cylinder instead of the framework, might be interesting for comparison to the new fvOptions approach. Of course, computational costs are higher.

[wrcase]

Hello Albrecht
It turns out if I let the simulation run longer (endTime = 20 s) I do get a rise in pressure and water begins to climb up the right wall. A look at the velocities shows that they are very slowly catching up to the wall velocity. Additionally the wall velocities at the first time step took on the value (0 0 -1) = r * omega in the -z direction, which tells me the setup and behavior are correct. It happened to respond much slower than I originally guessed.

Thank you for the suggestion, though. I may be able to apply what I learned from the testTubeMixer tutorial.

Kind regards,
Bill

[otm]

Looks like you simulated the quite classical Ekman spin-up time problem.
A quite good description can be found on this link.
http://go.owu.edu/~physics/StudentRe...asto/home.html

BR
Olle