[handicraft]

hi all
I am trying to simulate a car in wind tuunel using simplefoam (steady-state) solver (k-omega) without turbolence and i meshing geometry with SHM. final mesh had 4.4m cell. the problem is that residuals converge at high values (0.01). i do it with 1.4m cell and residuals converge at 0.0001 . i change the scheme to high order scheme (second order accurate) but nothing change.

What is wrong in my case? and how i can improve it?
Thank you very much for your help.





controlDict.zip

fvSchemes.zip

fvSolution.zip

0.zip

[arvindpj]

Try reducing the relaxation factors:

relaxationFactors
{

   field
   {
       p   0.8;
   }

    equations
    {
        U               0.2;
        k               0.5;
        omega       0.5;
    }
}

[handicraft]

Thanks for the answer,
i do this and here the residuals plot:

It's totally messed up

[arvindpj]

Could you share the full case with mesh?

[handicraft]

that means you need 0,constant,system folder? because they are huge (like 400 mbyte)

[arvindpj]

did you create the mesh using SHM or cfMesh.. then use foamCleanPolyMesh and send only the scripts.

[handicraft]

sorry I am new to OpenFoam and little confused. yes i'm using SHM, and when i use this command (foamCleanPolyMesh) nothing happend. only show "Cleaning ./constant/polyMesh"

[arvindpj]

Now the file size would be small and you could share.

[handicraft]

thanks, sorry but file still 1.4mb and forums don't let me upload here, so i uploded some were elese,
here the case:
http://www.mediafire.com/file/guewlz...3/car.zip/file

[pete20r2]

Try more cells on the ground around the car, there will be a velocity gradient under the both to the undertray and to the ground so you need a boundary layer of prism cells there too. Also try a cell size of about a quarter of what is currently under the car. For an easy test, turn on boundary cells for the whole floor patch and thicken the first refinement layer off the vehicle until it is affecting the floor around the vehicle.