Node merging
 

Hi Simon,

Thanks a lot for your so quick reply. I followed your advice for mesh smoothing and got it now.
Again I'll implement your recent suggestions. You are just amazing man, how do you manage to reply to all the posts.

Thanks again

Etesh

Assen
 

Hello my name is Assen! I have the same task, but I've never done a dynamic mesh.Would you give me a tutorial on your task.Тhank you very much in advance!
asen_aerodynamics@abv.bg

Time...
 

I am working on my MBA (At WPI) and I find CFD-Online to be a pleasant distraction. I can easily answer three posts in the time it would take me to formulate the Modigliani-Miller Proposition (with and without Taxes). I eventually get back to the school work, but the change is as good as a rest. ;^)

Most of the help is very short. just a few minutes and I can get something blocked out or just get someone on their way.

If I had more time, I would do more movies for the youtube site...

For kintata, check out www.youtube.com/ansysinc/

Hi Kintata,

Mesh Steps
ANSYS Version: 12.0

Open the tin file.


Create parts. To better set boundary conditions.

Concatenate curves of airfoil
Select all the segments of airfoil.


Set periodicity......Blocking--> Edit block--> periodic vertices (select pair of opposite nodes)



Create 2D block. blocking-->create block--> initialize block (2D planar if its 2D)



Point & edge association Blocking-->Associate--> vertex and edge-curve association



Split block blocking--> split block ( around the airfoil)


Point & edge association again



Create O-grid
Select the inner most block, create around block.

blocking--> create block--> O grid block




Delete block
Delete the inner most one (O grid block u just made).


edge curve association for O grid


Set meshing parameter blocking--> pre-mesh parameters



Pre-mesh


Check mesh quality, make sure no negative cells



Create mesh

Mesh check
No periodicity, and ignore single edges error message.

convert to unstructured mesh by right clicking on Premesh in Display tree


Output
Output solver: Fluent_V6

Set boundaries

Output mesh file

Assen
 

Thanks a lot! Would you tell me the steps in a fluent for dynamic mesh?

r u done wid the meshing thing?

Assen
 

I tried, but Failed to do it.Would you recommend a tutorial from which to start or something

I simply followed 3 videos uploaded by Simon and those proved to be of great assistance to me. www.youtube.com/ansysinc/

Looking for doing the same..
 

Hi all!
thanks to simon another time.. I'm trying to do a similar job as etesh, but before to insert my NACA0012 airfoil in a 3 circles structure, as the video etesh posted the attachement, I want to have the perfect hortogonality of the first layer over the airfoil.
I followed the Simon's videos, but I have 2 important differences, first, a blunt trailing edge, second, I need a O-grid cause my FF is circular.
I attach now 2 jpegs with my trailing edge mesh smoothed and my blocks structure.. Can someone of you tell me if is good enough and how to improve it?

Thanks all

3 suggestions...
 

You adapted it well to your situation (OGrid instead of a CGRID).

In the end, you have a very orthogonal mesh at the trailing edge... Depending on your solver, it may be better to relax the orthogonality and have a smoother transition between elements. To do that, make sure your trailing edge is in a unique part (such as TE), then recompute your premesh and convert to unstructured mesh (this is required to make sure the new part name is used). Then when you use the solver, you will find an option to "Release Orthogonality", Select the TE part.

Alternatively, you may prefer the mesh you would get by having a small round instead of a flat trailing edge. Then the orthogonality wouldn't be at the cost of a smooth transition. To make that change, just change the geometry, then associate the previous blocking file to the new curve, generate premesh, etc.

A third option would be to just use the mesh edit tools to manually move a few of these nodes around until you get what you want... A good compromise between orthogonality and smooth transition (cell area to cell area and angle to angle)

Mesh 2D in gambit
 

Hello everyone

I have to create mesh profile in 2D exactly like this image in gambit someone could help me please:(

http://co103w.col103.mail.live.com/a...CCAF4817F4F0||

Mesh 2D in gambit
 

Hello everyone

I have to create a mesh profile in 2D exactly like this please somebody could help me :(

http://co103w.col103.mail.live.com/a...CCAF4817F4F0||

Airfoil rotating w.r.t central axis
 

Hi,
I am trying to run simulation (2D) for 3 bladed NACA 0012 airfoil in vertical motion with respect to central axis using gambit and fluent. I know how to create mesh for single airfoil and run simulation on it, but i don't know how to create mesh for airfoil and to rotate it in fluent. I would be glad if anyone could help me out.
Here is a youtube video of what i actually want to do. Please help me out guys, its part of my thesis.
http://www.youtube.com/watch?v=1eyEtzdHDEE

Regards.

Slightly Different
 

Hi everyone,

I have a very similar problem to the one Etesh has outlined above.

I am currently trying to do a 2D model of a wind turbine blade, only my turbine is horizontal axis, rather than vertical axis.

This means that my blade profile isn't in the plane of the wind, so where Etesh designed a disk (3x120 degrees) with his blade profiles lying flat on the disk, I would have to design (I'm guessing here, I'm quite new to this CFD business) a kind of cylinder, or curved mesh. But this makes the problem 3D (I think).

Could anybody help me? I'm keen to do this problem in 2D as it as specific blade, for which I only have one profile, so any extrusions etc. would be all guess work.

Thanks,
Richie.

More similar than different
 

You could do the exact same as Etesh, except one of your 2D boundaries becomes the ground. Then extrude it by the length of your cylinder to get it to 3D.

Sounds like an odd design though, how would the air move the airfoils if it runs along them? perhaps I am not understanding your intent.

What I'm trying to model is a standard Vestas V52 wind turbine. But only a 2D profile of one blade.

If I turn Etash's design, making one side the ground the blade profile will be wrong.

I have been told that whereas an airfoil can be modelled with a mesh as per the video tutorials above, a wind turbine can't, because of it's circular motion. Is this correct?

All I want to do is model the airflow over the blade profile in 2D.

Thank's for replying so quickly.

Richie.

Oh, just a regular wind turbine then...
 

Oh, so you just want to model a regular wind turbine...

http://www.google.com/images?hl=en&s...=1906&bih=1043

You would need to model a 3D disk (or a 120 periodic degree segment with a single blade) with the rotating portion in it. The rotating portion would need to be a 3D model. My guess is that you would end up lofting the airfoil curve to create the geometry. In reality, you should have several airfoils and probably also twist along the blade (portions further out travel at higher speed and therefore should have a lower angle of attack or a different airfoil), then you would put that inside a larger box containing the tower and terrain. This last part is the only bit that is like Etesh...

For modeling the blade, look for other threads.

Here are a few pics of a wind turbine I did last year... I used MultiZone Hexa in this case, but you could use full hexa or tetra or what ever you are comfortable with. I think you first need to build or get your geometry. Note, it is just the mesh of a single blade... I would then rotate this...

Oh, just simple 2D airfoil...
 

Sorry, I got carried away... If you just want to model the 2D profile of one blade, then you just model that... You can do it just like my Youtube movie...

http://www.youtube.com/watch?v=tYrbScUH9RE

You would get the velocity of the blade by knowing the rotational speed of the wind turbine, the distance out the blade that the cross section is and the wind speed (vector addition).

It won't look as cool as the 2D model of the vertical axis turbine, but it is what you asked for.

Anything 3D with actual motion will require more effort, as in my previous post.

Thanks yeah, it really is the 2D modelling that is most suitable for me, I think.

I was thinking and hoping that the vector addition of the windspeed and bladespeed would be enough.

Which leads me on to my next question. I have followed those tutorials very closely, and they are responsible for 90% of my knowledge of ICEM (thanks), but, when I go to solve that mesh in Fluent, taking airflow to be incompressible, Fluent won't allow a Far Field Boundary, I think it would prefer an inlet outlet situation? Is this the case or have I wrongly assumed air to be incompressible? Also, I was looking at using the k-epsilon model of turbulence.

I usually create a tunnel myself... In that CGrid far field case I used a pressure far field. This would be a good question for the Fluent Solver Forum...