Centrifugal force?

[muraer]

Ok, thanks for the help! That should be going now

Best regards!
Mario

[gmontgomery]

I was able to use MATC to simulate a spinning disk. I used a formula in the MATC statement to make the angular velocity a function of time.

For the problem I eventually wish to solve the angular velocity will be a table such as

Time Omega (angular velocity)
0 100
1 300
2 600
. .

I think the best approach is a user function but am very confused by the documentation.

Has anyone done this?

Thank you.

George

[mzenker]

Hi,

there is a chapter "writing a user function" in the solver manual which explains a lot of things. Also a FORTAN 90 manual will help.
If you want to pass more than one variable to your user function, see here for instructions.

HTH,

Matthias

[Erick DNA]

Hi
i am studying the hydrodynamic behavior of a stirred tank and i have the angular velocity for boundary conditions of stirrer in the form:
Velocity 3 = Variable Coordinate 1; Real MATC "125*tx"
Velocity 1 = Variable Coordinate 3; Real MATC "-125*tx"
where 125 is the angular velocity...
the angular velocity is in degrees/s or in rad/s if all my units are in SI?

thanks

[annier]

Hi Eric,
in "125*tx",125 is a number and you can decide any unit to it. As you say that you are using SI unit, you will be using angular velocity as 125 rad/s so that it will give the resulting velocity in m/s.

Yours
Anil Kunwar

[schewitz]

Hi,

because of your awesome guidance I was able to calculate the stresses resulting from centrifugal force inside a running wheel of a radial compressor.
For validating these results I calculated disk rotating around the z-axis which has a constant thickness of 1mm and a bore in the middle. I calculated the stress distribution analytically on the one Hand and by FEM Simulation on the other.

For the centrifugal force I used the following Body force command:

Code: Select all

Body Force 1
Name = "BodyForce 1"
Stress Bodyforce 1 = Variable Coordinate 1
Real MATC "Density*omega^2*tx"
Stress Bodyforce 2 = Variable Coordinate 2
Real MATC "Density*omega^2*tx"
End

with Density=7850 kg/m^3 and omega=1047 rad/s


the FEM results looking qualitative quite plausible, but the deviations between the two cases are up to 23 N/m^2

Can anybody help me? I cannot finde the mistake.
The Excel sheet with my analytical calkulations, the Elmer project files and a screenshot of the vonminses stresses from the post processing Software ParaView are attached.

Best regards,
Matthias

[raback]

Hi schewitz

Probably there is no mistake if you're already this close.

Have you tried grid converge studies. I.e. increase mesh density parameter and study how the error behaves.

Are you using linear elements? You should probably get better accuracy with the with the same resources using quadratic elements (or p-elements).

You could use SaveLine with "Polyline Coordinates" to plot the stresses through a line for comparision with the analytical solution.

-Peter

[schewitz]

Hi Peter,

thanks for your quick Response.
I did also some computations with much higher mesh densities. The results are exactly the same. But I will try to use other elements.

Matthias

[mzenker]

Hi Mario,

I can't get him to add the term "variable coordinate" in the .sif file, but editing the .sif in the GUI also seems to work.

you can use a free text field to add your "variable coordinate" term. This has the advantage that it will be saved in the ElmerGUI project. If you edit the sif file manually, it will be overwritten on next sif file generation by the GUI.

HTH,

Matthias