[ashkan]

Hi All,
I was wondering if anyone have experience with overset solvers of foam-extend-4.0. I noted that the foam-extend overset library performs much superior both in terms of accuracy and speed compared to OpenFoam v1706.

However, it seems that it cannot handle when there are regions with no overlaps, similar to the attached snapshot.

Hereby, I was wondering if anyone had experience with foam-extend-4.0 overset solver that can either confirm my understanding or help me on how to setup such cases.

I really appreciate any inputs.

Ashkan

[Santiago]

Ashkan, where did you get overset for foam-extend?

Id like to give it a try and compare against the official version.

[ashkan]

Originally Posted by Santiago

Ashkan, where did you get overset for foam-extend?

Id like to give it a try and compare against the official version.

Hi Santiago,
I followed the information in this link

http://foam-extend.fsb.hr/blog/2017/...n-foam-extend/

Basically, you need to email Prof. Jasak to get the link to download the Overset library (assuming you have already compiled foam-extend 4.0) and then just compile the overset library.

Kind regards
Ashkan

[Yann]

Hi Ashkan,

I ran a similar case in OpenFoam v1706 and I have issues with inconsistent values in the vicinity of the wall and the part of the moving mesh being outside the background mesh.

What's the behaviour of your simulation when running such case with foam-extend-4.0 and what's your understanding about it?

Regards,
Yann

[christoph45]

Hi
I'm doing a test case based on the "simpleRotor" tutorial from OFv1706 and I changed it to an active valve scenario.

Check out setup.jpg. It's a valve (yellow - "hole" - type wall) with the surrounding mesh (grey - type overset)

So at t=0 the valve is open and the overset mesh still inside the background mesh. See valve_open.jpg for cellTypes and v vectors. (red=hole, grey=interpolated, blue= calculated)

As the valve closes the cellTypes change, as I would say, the right manner. - They change to red. - But still when you look closely, there's still a non zero solution in the cellTypes which shouldn't been calculated.

What do you think about this? Am I expecting the right thing? Or am I demanding something openfoam can't handle? Like moving the overset mesh outside of the background mesh?

Also why aren't all cells around the "hole" (cellType 2 - red) being interpolated (cellType 1 - grey), but cellType 0?

Chris

[ashkan]

Originally Posted by christoph45

Hi
I'm doing a test case based on the "simpleRotor" tutorial from OFv1706 and I changed it to an active valve scenario.

Check out setup.jpg. It's a valve (yellow - "hole" - type wall) with the surrounding mesh (grey - type overset)

So at t=0 the valve is open and the overset mesh still inside the background mesh. See valve_open.jpg for cellTypes and v vectors. (red=hole, grey=interpolated, blue= calculated)

As the valve closes the cellTypes change, as I would say, the right manner. - They change to red. - But still when you look closely, there's still a non zero solution in the cellTypes which shouldn't been calculated.

What do you think about this? Am I expecting the right thing? Or am I demanding something openfoam can't handle? Like moving the overset mesh outside of the background mesh?

Also why aren't all cells around the "hole" (cellType 2 - red) being interpolated (cellType 1 - grey), but cellType 0?

Chris

Hi Christoph,
I have seen the same issue in my simulation when used 1706. However, I have switched to 1712 (released on 30 December 2017) and I do not have non-zero solutions in the outer domains.

Also, your case should be modeled very well with foam-extend overset solver as well.

Ashkan

[christoph45]

Thanks a lot for the hint, Ashkan. Works now!

[amanbearpig]

As there's already an active topic on the foam-extend overset mesh method, I thought I'd jump in with a quick question about it. I've used the overset mesh method with OpenFOAM-v1706+ but not the foam-extend version, does the foam-extend version work easily with 6DoF problems like the v1706/v1712 versions do? I'm not as experienced with foam-extend.

[skyoung]

Hi everyone,

I am using Probes in controlDict to get the pressure on the moving wall. I have set the fixedLocations false, but the presure of the probes doesn't change during the simulation. I guess the Probes just are fixed in the backgroud mesh and they are identified in the hole mesh, they don't move with the Ovserset mesh. Do you have any advice to modify this setting?
I really appreciate your help.

Best regards,
Yang

[mcaus]

Hi Ashkan and Christoph,

I'm setting up a case very similar to yours, with a part of the overset mesh leaving the background mesh domain. I've tried with v1806 and foam-extend4.1, but in both cases, I get unphysical flow over the no-slip boundary where the overset mesh leaves the background mesh.
Could you explain how to set up oversetMeshDict (for foam-extend) or eventually share your case?
Christopher, did you actually get your case working on v1712? As far as I know, the openfoam.com versions do not support this specific functionality, would be happy to hear otherwise!