[joshcouch]

Hi,

I've been having trouble with my simulation of a dfbi body in a two phase (water and air) VoF simulation on Star.
The body is subject to a downwards step force applied for 0.5s.

I am using an overset mesh. The problem is that the free surface is very unstable and drops, residuals are high (1-100) and I can't seem to fix this, no matter the time step I use. I have tried a time step as small as 5e-5 and 20 inner iterations but the problem persists. The maximum velocity I have observed is 100m/s and the smallest cell is 0.075m.
Has anyone had a similar problem? Or can anyone offer some suggestions?
Many thanks in advance,
Josh

[dmirel]

Hello,

It might be hard to capture a free-surface movement with 75mm mesh cells...
There is a tutorial in the manual DFBI- Lifeboat with Overset Mesh. Maybe you can extract some valuable information from it.

Demirel

[joshcouch]

Hi Demirel,

Thanks for your response. The 75mm cells are because I am anticipating waves of roughly 1.5m height, and for VoF waves Siemens recommend 1/20 of the wave height, perhaps this guideline doesn't work for 'self-generated' waves (as opposed to VoF wave model). It also recommends 1/80 to 1/100 of the wavelength for non-vertical directions, with gives approximately 1.25m. I'll try refining it in these directions, but the run time may end up being too long. I'm using a trimmed mesh with anisotropic mesh sizes by the way.
I have also done the lifeboat tutorial, and they use a more coarse mesh than I do, and the default values also produce large residuals (~100) for momentum.
Any further help would be much appreciated,
Josh

[joshcouch]

Originally Posted by Steklo_Plastik

If the body falls out, then it is necessary to reduce the step. If a free surface falls out, then you need to correctly build an overset, the intersection with the grid of the medium section and the background.
Volumetric control of only the section of media from the background and overset.
The correct intersection of the offset with waves or a free surface is important.
At the intersection, the cells must be the same.
6 million cells.
Attachment 75543
Attachment 75544
https://youtu.be/fcg2xHbiF9c

In the settings there is a speed limit of 20 m/s. it is Necessary to change the speed limit .

Hi Steklo,
Thanks for your reply. From what you've said, my overset mesh is likely to be the problem, as my free surface drops within the overset region. I have attached a picture of this and my mesh.
I only seem to have the one mesh continuum, is this because I have used automated meshing and you have used region based? Should I tick the box that says 'per-part meshing' on my automated meshes?
Thanks,

Josh

[joshcouch]

Originally Posted by Steklo_Plastik

I do not use automatic grids. A lot of extra cells, a lot.
You need two areas with your own grid.
Paul cells in one step.
Use DFBI free movement and rotation.
Attachment 75551
Attachment 75552
The study should have a natural accompaniment. It should be like a natural phenomenon. My principle.

We must be either English or Russian. Only in this way can we understand each other. Probably.

Another example for you.
https://youtu.be/YrBHhoUfpHM
https://youtu.be/DecU0tiKEOw
https://youtu.be/eZ-N1SpKRU0

It is not enough to be a good engineer, physicist and mathematician. We must love and understand nature, life itself.
Only then can positive results be obtained. Or stay blind for life. Take beer and dried fish, go to the shore of the sea, lakes, rivers and look at them carefully. Hug them immense and they will tell you about themselves.

Hi again Steklo,
Thanks for those screenshots and clips, they have helped me adjust my mesh slightly and after also changing the time order to first (with multistep) I seem to be getting some good results!
Josh