Wall Forces at cylindrical structures

[Anonymous]

Hi,

as I am writing my master thesis I am also working with Autodesk CFD. I want to get some surface results around cylindrical structures in air at different Reynolds-numbers (keyword: karman vortex street). I run simulations with a 2D cylinder (d=30mm), transient, sst k_omega with time steps around 0.0002s. 

For the c_p values (pressure value in the front) the results seem to be alright. Now I would like to measure c_d/c_f (force value at the whole structure in one direction). I tried the Wall Calculator (for Fx) and I'm also working with the Flags "Wall_force_History" and "Resid_bdry_force_calc" to get the wall function over time (to get Fy(t)). Unfortunately the results don't match (for the last time step). I also tried different simulations.

My questions:

1. What is the difference and which are correct?
2. Is there another way to calculate the forces on the structure?
3. Can I somehow get the flag-files to make modifications or write my own? I can't find them in the Autodesk CFD folder.

Thank you very much

[Jon.Wilde]

Hello,

Looks like an interesting project, feel free to share some images or videos once you have it run out!

Could you re-ask your first question? Not sure what you mean.

The Wall Calculator should be OK here I think

I don't think you can modify the flags but you could use the API, it depends what you are looking to achieve though?

Thanks,

Jon

[Anonymous]

Hi

to the first question: the results of wall_force_history (.csv file in project folder, with flags "Wall_force_History" and "Resid_bdry_force_calc") and the results of the Wall Calculator in the programm itselts differ in some simulations. But I am working on this, maybe I have to refine some meshing, timesteps or the like. 

With the two mentioned flags I get the csv-file but there are empty rows and text in it. I would be nicer to just have the values over time to look at it in excel. API might be a solution, but it would be easy just to chance the wall_force_history flag.

If I get so good and reliable results I will post them.

Thanks

[Anonymous]

So I did some more simulations to find out if the wall-forces of the wallCalculator (last step) and of the.csv-file (last entry), written by the flags "Wall_force_History" and "Resid_bdry_force_calc" are alike. It seems that the reynoldsnumber has an impact. With Re=10^7 the results are the same - for Re<10^5 the results differ up to a factor 20. 

My simulations parameters are:

Circulare shape (diameter: 3mm, 30mm, 3000mm)

Inlet: u = 1 - 5000m/s, outlet: pressure, sides: slip

Mesh Refinement at the structure - meshsize 2mm - 0.2mm

No. of Layer 10-15, Layer enhancement 1.5

ADV 5 

Turbulence model: k-omega SST

Steady state

Around 2500 iterations

Simulation, which did have the same results:

D = 30 mm, u = 5000 m/s, mesh: 2mm-0.2mm     --> Re = 10^7

D = 3000 mm, u = 50 m/s, mesh: 2mm - 0.2mm     --> Re = 10^7

D = 30 mm, u = 10 m/s, mesh: 2mm BUT k-omega SAS     --> Re < 10^5

Simulation, which did NOT have the same results

D = 3mm, u = 50 m/s, mesh: 2mm-0.2mm     --> Re = 10^5

D = 30mm, u = 50 m/s, mesh: 2mm-0.2mm     --> Re = 10^5

D = 30mm, u = 1, 5, 10, 20, 30 m/s, mesh: 2mm     --> Re < 10^5

Till now I haven´t done any simulations with around Re=10^6... 

Do I still have any mistake in my simulations? How do the flags "Wall_force_History" and "Resid_bdry_force_calc" calculate the WallForces?

[Anonymous]

And I did some more simulations - same setup but different turbulence models. 

Results:

The k-omega models RC (both) and the k-omega with the flag "sst_new_iwf" gives with all re-numbers the same results for the WallCalculator and the WallForceHistory in the same simulation. The models k-omega, k-omega DES and k-omega SAS just have the same results with really high re-numbers.

This brings me back to my question:

What it the difference between getting WallForces of the WallCalculator and the WallForceHistory in the same simulation??! 

As i would think the WallForceHistory just recodes the results getting of the WallCalculator, BUT this seems not to be the case!