[jango]

Hello Everybody,

To get used to OpenFOAM cases and solvers I`m playing around a bit with the karman vortex street. With the icoFoam Solver for laminar flow at low Reynolds nubmer this isn`t a problem.

But now I want to increase the Reynolds number und add a k-epsilon (near) wall treatment. The fluid is air with a temperature of 273 K and the wall temperature of the cylinder is 573 K. For this case I assumed, that rhoPimpleFoam should be the best solver. With the standard adjustments for this solver, the case runs without any convergence problems but I can see no vortexes. So I changed the runtime for the turbulent case an adjusted the material properties (thermophysical properties) to constantTransport with the according values of Pr and mu for air.

After a few iterations the solution diverges. Now I have a few questions:

Is the rhoPimpleFoam solver the best choice for this case ?
Adjusting the thermophysicalProperties according to this Problem (considering adjustments for Temperature and material properties for air) I get errormessages and convergence problems. Should I adjust the underrelaxation factors to stabilize the solution ?

Thank you a lot for your Help

Yours,

Jan

[Bernhard]

Do you understand turbulence modelling? What do you mean by 'k-epsilon wall treatment'? If you're interesting in vortex shedding then you can better use a large eddy simulations, then you're solving for at least a part of the turbulent spectrum. And why do you include temperature? This is not needed for the vortex street to develop. Start with isothermal simulations. If you're sure those are correct, then you can start with solving for non-constant densities. Good luck!

[jango]

Hello Bernhard,

thank you for your quick response. Of course the temperature isn`t important for the vortex street. As I learnded CFD with Fluent at the university, I have various test cases based on a fluent script and I'm just trying various settings and solvers for the better understanding of OpenFOAM.

Thank you for the hint with the large Eddy- Simulation, but is rhoPimpleFoam the best choice for the turbulence modelling (and maybe heat transfer) or is simpleFOAM the better choice ?

[Bernhard]

simpleFoam is designed for steady-state problems, and your case is transient by definition, so simpleFoam is not an option.

[jango]

Thank you for the hints, that helped me a lot. So I just successfully calculated the karman vortex street with the icoFoam solver, increased the reynolds number for the turbulent case and calculated it with the psioFoam solver. Then I changed to compressible medium (air) with the rhoPimpleFoam solver. Then I mapped the fields from this calculation adding Temperature and running the job again. But evaluating the results, I can see, that temperature is not considered in my calculation. Could it be, that my initial definitions in the files 0/T and 0/alphat are erroneous ?

My setting for T:

dimensions [0 0 0 1 0 0 0];

internalField uniform 293;
internalField2 uniform 593;

boundaryField
{
far_field
{
type symmetryPlane;
}
outlet
{
type calculated;
// inletValue uniform 293;
// value uniform 293;
}
inlet
{
type fixedValue;
value uniform 293;
}
cyl_wall
{
type fixedValue;
value $internalField2
}
frontAndBackPlanes
{
type empty;
}
}

and alphat:

dimensions [1 -1 -1 0 0 0 0];

internalField uniform 0;

boundaryField
{
far_field
{
type symmetryPlane;
}
outlet
{
type calculated;
value uniform 0;
}
inlet
{
type calculated;
value uniform 0;
}
cyl_wall
{
type alphatWallFunction;
value uniform 0;
}
frontAndBackPlanes
{
type empty;
}
}

What can I do to calculate the temperature ?

[charbel]

Originally Posted by jango

What can I do to calculate the temperature ?

See this link to learn how to add temperature to icoFoam solver.

[jango]

Thank you charbel, I think that could be the solution. But I'm still a bit confused. Taking a look at the rhoPimpleFoam solver for compressible fluids and the angledDuct tutorial case, there`s already temperature added in this tutorial case and the field must already be there, but in my case, the fluid has no temperature. As the fluid should be heated by the temperature applied to the cylinder, could it be that my initial conditions for alphat are wrong ?