[kimou]

Hi Guys,

I am a new Openfoam user, and i am currently working on flow around an airfoil profile (externel flow). I work with overPimpleDyMFoam (transient regime), my mesh is 100k cells (using overSet Mesh) my problem is that my simulation takes a lot of time, to reach a duration of 0.2 second, my computer calculates during 2 days (it's huge time)

knowing that my current number maxCO = 8, and y + = 10.


Please help me !!!!!!!! please, !!!!!!

[Swift]

Hi Hamdani,

I guess the obvious question is what computer do you have?

Are you running in parallel and what commands are you using to run in parallel? I ask this only because you say your are new to OpenFoam. I don't mean to offend you.

Assuming your computer is able to handle CFD, could you upload the following:
1) A screenshot of your case so that we can see what your are doing
2) A screen shot showing your mesh.
3) fvSchemes and fvSolution files.
4) A snippet of the output from your log file when running overPimpleDyMFoam.
5) Output from checkMesh,


Also what openfoam version are you using. I think there have been great improvements to overset in 1812.

With this information, you might get more help.

Thomas

[kimou]

Hi Thomas,

first of all I want to thank you for answering me.
You will find attached "systeme folder "




- yes I use the version v1812 which is compatible with oversetMesh
- My computer : Intel xeon 12 cores, RAM 32GB,

- Yes i'm running the simulation in parallel with :

----------------------------------

numberOfSubdomains 10;

method hierarchical;

coeffs
{
n (5 1 2);
delta 0.001; // default=0.001
//order xyz; // default=xzy
}
----------------------------------------








system.zip

cfd_online2.jpg

cfd_Online1.png

checkMesh.jpg

solver.jpg


Tthanks you Thomas


Hakim hamdani

[Swift]

Hi Hamdani,

I see you have duplicated your question. I answered there as well. I thought I was going crazy. This was my answer there. I will no longer answer on that post.

As I had mentioned, you need a minimum of four cells from your airfoil wall to the edge of your overset mesh. I can't count how many you have, but it looks to be many more. I would first start by reducing the number of cells in overset. This should substantially improve your simulation time. Here is a link for overset best practices. Its not for OF, but the principles should still apply.

Then you need to get your background mesh to be the same size where your overset mesh is located. This will make it easier for OF to get convergence in these regions, which might help the fact that you are not converging over 10 iterations which is why your simulation may be taking so much longer.

I would set your tolerance for omega a little lower or set the minimum iterations to at least 1. The reason for this is that the value is not changing because it is starting with a very small value. I have seen in my own work that by reducing the tolerance on these values can help if no iterations are being performed.

Regards,
Thomas

[kimou]

Hi Thomas,

I apologize for the confusion by posting publications.
Thank you Thomas, your information is gold

regrets,
Hakim