[Pol.P]

Hello,
I am an Engineering student and a new member of CFD-Online.
I am currently doing an internship to evaluate the utility of the CFD approach to improve paragliding wings. I use ANSYS Workbench (DM, Mesher, Fluent) to perform my CFD simulations.


I would like to simulate the comportment of an airfoil in 2D for several incidences (3° to 15°). That is why I propose the domain (figure_domain).I have imported my 2D domain from SolidWorks then I modify it in DesignModeler (Creation of line and projection to slice the domain) to allow me to refine the mesh.


The meshing step (using of Mapped face meshing and size meshing) seems good but when I implement it in fluent and try to solve the problem all the residuals diverge.
I have already followed many tutorials that show how to perform airfoil simulation but no one is relevant about what I am looking for.


I think it is a problem of geometry (something is wrong when I create additional geometry to allow me to refine the mesh) because when I resolve the same problem with a simple mesh (Inflation + Size defined on the airfoil) it is working well.


Fluent parameters : k-epsilon method

If someone can take a look at my project I would be very grateful.


Best regards,


Paul

[FJSJ]

Hi Paul,

I am not sure if you have a problem with your geometry/mesh or with your setup in Fluent. What tipe of mesh are you using? What about your quality mesh criteria? Skewness/Aspect Ratio/ Orthogonal Quality? Some warnings from Fluent "check case"?

[Pol.P]

Hi,

Thanks you so much for your interest.

I checked criteria that you talked about :
ANSYS MESHER > Mesh statistics
- Quadrilaterals Mesh
- Aspect Ratio :
min : 1,0004
max : 44,355
average : 5,577838
standard deviation : 4,351

FLUENT > Report Quality :
- Orthogonal quality 0,30582
- No warnings from fluent "check case"

I attached a picture of my mesh which show some geometrical quality characteristics

[Pol.P]

Moreover,

As I say above I would suspect a geometry/mesh problem that imply some undesirable effects during the solving.
To argue my words I attached the evolution of residuals (Mainly : divergence of continuity).

Thanks Javi if you can give me some help.

If you need some others informations, I can share the project.

Regards

[FJSJ]

Ok. If you want to share your project, I can have a look!

[Pol.P]

Hi Javi,

Here another link to access the project.

https://www.dropbox.com/sh/khk369onn...N1GfTUc3a?dl=0

Thank you.

[antm324]

Hi Paul,

I am no means an expert on CFD but i will try to help with what i can,

what is your y+ values?
why Have you chosen K-epsilon as your turbulence model?
What is your Reynold number?

Anthony

[Pol.P]

Hello Anthony,

I calculated the Y+ with the CFD-online tool and it is equal more or less to 2.5e-5 m.
The Reynolds number is equal to 2.2e+6.

I choose K-epsilon because it is an usual method to calculate the flow with turbulence.

I follow a tutorial of my school fluid teacher (flow over a NACA airfoil using k-epsilon and the same mesh type and boundaries conditions that I have implemented) but we had not make the mesh. That is why I think that is a problem of meshing/geometry definition.

Paul

Originally Posted by antm324

Hi Paul,

I am no means an expert on CFD but i will try to help with what i can,

what is your y+ values?
why Have you chosen K-epsilon as your turbulence model?
What is your Reynold number?

Anthony

[antm324]

When you say you have calculated your y+ here, what y+ does your mesh give you when you check it from the simulation?

In my little experience I found k-epsilon rather poor at predicting results so it maybe worth looking T the different choices in turbulence model as well ensuring that your y+ is appropriate for such a tour opulence model

[hoyos98]

I suggest you visit https://aeroptimal.com/mesh (you must create an account to use this module), where you can create a full structured airfoil mesh - https://youtu.be/4Opu0zk7gFk . You can export .su2 .msh .foam .vtk