[fredicenci]

Hello guys. I am new using openFoam and i am trying to simulate a turbulent flow around a circular cylinder with 1 meter of diameter and 5 of length, the cylinder is fixed on a wall. I would really appreciate some help =).
Well, after around 1000 iterations the drag coefficient start to raise to infinity, and it should converge to around 0.75. Also, the lift coefficient goes crazy. I did this simulation using ansys and i am trying to reproduce it using openfoam.


Some more information: k and omega was calculated according to these equations. https://www.cfd-online.com/Wiki/Turb...ary_conditions


also the nut was calculated with the viscosity ratio of nut/nu = 10.




I am not really sure about the boundary conditions. If someone could take look in my files and give some tips why my simulation is not working i would really appreciate.


link with the files compressed : https://www.dropbox.com/s/nmjdvhjypp...on.tar.gz?dl=0



Thank you guys!

[jeytsav]

Dear fredicenci

I am new to openFoam and CFD in general but it's been a couple of months that I am dealing with a simple case.

I reviewed your files and I would like state some things


in the 0/U velocity BCs try using fixedValue 0 instead of noSlip at walls. I can not really tell the difference between the two, but for my cases I am using 0 velocity at the walls as listed below.

    cylinder
    {
        type            fixedValue;
        value           uniform (0 0 0);
    }

    bottom
    {
        type            fixedValue;
        value           uniform (0 0 0);
    }

in the 0/p presuures BCs, try using zeroGradient at inlet instead of fixedValue zero. Since you are already defining the velocity at the inlet, you don't want to define the pressure as well.

inlet
    {
        type 		zeroGradient;
    }

In the constant/polyMesh/boundary, you are using symmetryPlan for the symmetries of your domain which I think it is not a valid boundary. Use symmetryPlane instead or better use symmetry (which is what I am using for my cases)


Regarding the turbulence properties you are using I am not very familiar with this yet. Maybe you can try solving as a laminar case first to check whether the turbulence setting is the one to blame for the simulation failure.

Could you also provide the postProcessing results in order to have a better picture of the simulation.?

Kind regards

[fredicenci]

Originally Posted by jeytsav

Dear fredicenci

I am new to openFoam and CFD in general but it's been a couple of months that I am dealing with a simple case.

I reviewed your files and I would like state some things


in the 0/U velocity BCs try using fixedValue 0 instead of noSlip at walls. I can not really tell the difference between the two, but for my cases I am using 0 velocity at the walls as listed below.

    cylinder
    {
        type            fixedValue;
        value           uniform (0 0 0);
    }

    bottom
    {
        type            fixedValue;
        value           uniform (0 0 0);
    }

in the 0/p presuures BCs, try using zeroGradient at inlet instead of fixedValue zero. Since you are already defining the velocity at the inlet, you don't want to define the pressure as well.

inlet
    {
        type 		zeroGradient;
    }

In the constant/polyMesh/boundary, you are using symmetryPlan for the symmetries of your domain which I think it is not a valid boundary. Use symmetryPlane instead or better use symmetry (which is what I am using for my cases)


Regarding the turbulence properties you are using I am not very familiar with this yet. Maybe you can try solving as a laminar case first to check whether the turbulence setting is the one to blame for the simulation failure.

Could you also provide the postProcessing results in order to have a better picture of the simulation.?

Kind regards

Dear Jeytsav,


I really appreciate your help, I took your advice and I did the changes for the symmetry boundaries (I used symmetry), for the velocity on the walls (uniform (0 0 0)) and the pressure inlet ( I used zeroGradient) . Also i had to change the 0/P and 0/U for the symmetry boundaries from zeroGradient to symmetry.


You can get all the files of my simulation on this link: (Also the log FILE)


https://www.dropbox.com/s/rt4x6yz24y...AR.tar.gz?dl=0


I ran the simulation for the laminar case and came out the following error after the time 1.2.


#0 Foam::errorrrintStack(Foam::Ostream&) at ??:?
#1 Foam::sigFpe::sigHandler(int) at ??:?
#2 ? in "/lib/x86_64-linux-gnu/libc.so.6"
#3 Foam::GAMGSolver::scale(Foam::Field<double>&, Foam::Field<double>&, Foam::lduMatrix const&, Foam::FieldField<Foam::Field, double> const&, Foam::UPtrList<Foam::lduInterfaceField const> const&, Foam::Field<double> const&, unsigned char) const at ??:?
#4 Foam::GAMGSolver::Vcycle(Foam::PtrList<Foam::lduMa trix::smoother> const&, Foam::Field<double>&, Foam::Field<double> const&, Foam::Field<double>&, Foam::Field<double>&, Foam::Field<double>&, Foam::Field<double>&, Foam::Field<double>&, Foam::PtrList<Foam::Field<double> >&, Foam::PtrList<Foam::Field<double> >&, unsigned char) const at ??:?
#5 Foam::GAMGSolver::solve(Foam::Field<double>&, Foam::Field<double> const&, unsigned char) const at ??:?
#6 Foam::fvMatrix<double>::solveSegregated(Foam::dict ionary const&) at ??:?
#7 Foam::fvMatrix<double>::solve(Foam::dictionary const&) at ??:?
#8 Foam::fvMatrix<double>::solve() at ??:?
#9 ? at ??:?
#10 __libc_start_main in "/lib/x86_64-linux-gnu/libc.so.6"
#11 ? at ??:?
Floating point exception (core dumped)



So, what do you believe the problem is? Maybe the mesh? I meshed it on Salome and imported as UNV file to openFom. When I use the command checkMesh it returns Mesh OK.


Thanks again!

[jeytsav]

Dear fredicenci

It seems like your simulation is exploding due to velocity.

I tried to reduce the relaxation factor of velocity from 1 to 0.7 into the system/fvSolution file.

I solved for SteadyState and the simulation seemed to oscillate around Cd = 1.85 (After 1000 iterations) which is probably not what you were expecting as you mentioned.

I also switched to CrackNicolson (exactly like the case you uploaded) and just changed the relaxation factor as mentioned above. The simulation seems to be following the same behavior.

You can try running it for much more iterations to see if something changes.

Switching to turbulent flow we can expect an increase in drag.

I have not tested the turbulent case of yours, let me know if there are any further problems.

Questions: Why do you expect a Cd of 0.75? Do you have any experimental results? Or it is just a prediction ?




Some tips

You can sacrifice some accuracy in order to get faster results by changing the relTol setting into the solvers from 0 to 0.01. In my opinion this will lead to faster solution without affecting that much the accuracy. You can try on your own and see what happens.

You can also try using potentialFoam before running simpleFoam in order to achieve faster solution and decrease the chance of simulation explosion. It seems that in this small mesh, the use of potentialFoam is not that necessary but in larger meshes (in terms of cell count) and finer meshes, maybe its needed.

You can also try adding into the system/fvSolution file, some residual controls in order your simulation automatically stops when the residuals reach below the set value.

Residual Controls

SIMPLE
{
    nNonOrthogonalCorrectors 1;
    residualControl
    {
        p               1e-5;
        U               1e-5;
	k         	1e-5;
	omega	1e-5;
    }
}

Kind regards

[fredicenci]

Dear Jeytsav,

I was reducing the relaxation factors and i was getting the results just like you. I got the Drag value of 0.75 from experimental experiments, and i have already gotten the same drag coefficient value using ansys. Well, i will try run more iterations. Thanks for the tips to achieve faster solutions.

You helped me a lot already, I really appreciate. I will post further results. =)

[jeytsav]

Dear fredicenci

No problem!
I wish you good luck with the simulation! Please let me know if you manage to get the correct results and what settings did you finally use!

Kind regards