CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   ANSYS Meshing & Geometry (http://www.cfd-online.com/Forums/ansys-meshing/)
-   -   [ICEM] Problems with meshing of wind turbine (http://www.cfd-online.com/Forums/ansys-meshing/108208-problems-meshing-wind-turbine.html)

Dosis October 17, 2012 05:29

Problems with meshing of wind turbine
 
2 Attachment(s)
Hey guys,

I am pretty new to Ansys ICEM and I am currently working on meshing a whole horizontal wind turbine (50 m radius) with tower and ground. At the end I want to compare Ansys CFX and OpenFOAM based on this mesh.

My plan is to have at the end a stationary far field domain with the tower and a rotating disk/domain for the rotor.

Up to now I managed to mesh the rotor disk (14 million cells, hex) with a sufficient quality. As the blocking of the rotor is a lot of work, I used the symmetry of the rotor to save some work. So I meshed only one third of the rotor and then copied the geometry and the blocking of one blade.

This worked so far, now I have the problem to include the far field and the tower. To use the symmetry of the rotor I first had to exclude the farfield (not symmetric like rotor) and the tower. Now after having the rotor blocked and I would like to "extend" my blocking to the whole domain, but I am not getting this done correctly. And I dont know if its possible anyway..

So my questions: Is it somehow possible to create the complete meshing in one ICEM project?

I attached some screenshots so that you guys know what I am talking about, the tower is not included yet. The size of the far field is also not final yet.

I would be very grateful if you could help me with my problem!

Thanks a lot in advance!

Regards from Germany,

Bastian

diamondx October 17, 2012 11:41

so you did a blocking just for your wind turbine... the only way to expand it is to use the extrude faces... otherwise you should have started with a initial block around the hole geometry them perform splits and use index blocking to ease the works. i can't think of another way to expand you blocking, may be a new blocking of farfield then a merge... i don't think that will work.
I hope i'm wrong, let's see opinions of other member...

9021907902 October 17, 2012 22:30

hi...
I'm new to icem cfd.I'm doing analysis on an aircraft model as my final year project.I would like someone to help me in meshing the model.Actually when I meshed the model, I'm not able to capture the leading edge properly...Plz help me with this....

Dosis October 18, 2012 03:36

yes, I did the blocking only for the turbine as I otherwise would have not been able to use the symmetry of the rotor. Blocking (and meshing) farfield and turbine separately is maybe a better option, I already thought about that. To merge them in ICEM did not work so far.

I think in OpenFOAM there is a possibility to use two different meshes and as far as I know in CFX as well.

As the rotor disk should rotate I anyway need the GGI interfaces right? But I am not sure about the influence of the mesh cell ratio between both sides of the interfaces on the quality of the results. Is there a "rule" for a minimum cell ratio between the rotor disk side and the farfield side of the interfaces?

Thanks a lot for your help!

regards from Germany,

bastian

energy382 October 18, 2012 07:08

Quote:

Originally Posted by Dosis (Post 387248)
yes, I did the blocking only for the turbine as I otherwise would have not been able to use the symmetry of the rotor. Blocking (and meshing) farfield and turbine separately is maybe a better option, I already thought about that. To merge them in ICEM did not work so far.

I think in OpenFOAM there is a possibility to use two different meshes and as far as I know in CFX as well.

As the rotor disk should rotate I anyway need the GGI interfaces right? But I am not sure about the influence of the mesh cell ratio between both sides of the interfaces on the quality of the results. Is there a "rule" for a minimum cell ratio between the rotor disk side and the farfield side of the interfaces?

Thanks a lot for your help!

regards from Germany,

bastian


yes, it's possible to do it in one project. but you've to use sub-topo to stay on top of things. just rmb on topology -> create sub-topo -> select blocks to extract. you can either merge them together later on, or do meshing for each sub-topo and connect them in cfx-pre via ggi.

I prefer to merge the sub-blocking and do the interface in ICEM.....but hey, that's up to you.

you need 3 rot/stat interfaces. inlet, outlet and circumferential direction (as you don't have a casing around your wind turbine :D).

what shape does the tower have?

the biggest benefit of using sub-blocking is, that you have a consistent mesh and no cell size issues @interfaces.

but if it's not possible to get a good mesh (e.g bad cell angles), rather use a ggi or use a hybrid mesh.

I can't give you further advice without any pics of tower shape etc.

Far November 11, 2012 01:30

hey energy382

Can you give more details about the sub-topo. Where to use and how?

BrolY November 12, 2012 03:32

What about making the blocking of your farfield without the turbine, conver to unstructured mesh, and do a conformal merge wih the unstructured mesh of the turbine.
That's the easiest (and fastest) way of doing it I think.

Far November 12, 2012 03:41

@Broly What about making the hybrid mesh : Hexa in vicinity of blade and tetra in the rest of the domain? Then you may connect the both regions through merge or interface

But I would really be interested in Hexa meshing

BrolY November 12, 2012 12:20

Yep it should work as well, but I thought Hexa was the only option for this problem ;)

diamondx November 12, 2012 13:10

where is the geometry ?

Far November 12, 2012 13:13

here.....................


https://dl.dropbox.com/u/68746918/Windturbine.tin

diamondx November 12, 2012 13:14

link doesn't work, can you zip it ? make sure its in public folder. thanks a lot

Far November 12, 2012 13:22

working fine

Dosis November 13, 2012 05:00

hey guys,

@ energy382 sorry that I didn't reply on your post but I was on vacation and had some other stuff to do

now I am back on my mesh... for now I meshed farfield and rotordisk separately (hex) and I am now testing GGI connections to connect both meshes. I also worked a bit on your hint with the sub-topology, although it is difficult to use the periodicy of the rotor, this might work. I have to get back to this if my actual concept does not work.

energy382 November 13, 2012 06:41

I've done this kind of simulation before (with ICEM hexa). As soon as time permits, I'll give you some suggestions


All times are GMT -4. The time now is 09:18.