Blade meshing

Johannes,
I cannot add anything different from what you have stated above... I came to almost the same conclusions as yours as you can infer from the attached file (surely much more messed up than your tidy one). I was just about to give up and beginning to think to put hands on python code to create a tapered-twisted blade when I realized that I needed penthaedra rather then hexahedra. Performing an extrusion which leads to some mumbling on the part of SMESH but adding sub-meshes on root and tip (as you taught me in the previous messages) did the trick.
Next step will be compounding the two meshes got with the long sought after procedure you suggested.

Bis zum nächsten Mal,
Pier

Link:
pier.unirc.eu/cestino/blade2_work.hdf.gz

Post edited by: Pierpaolo, at: 2009/03/28 00:00

... what worries me now is the exclamation mark on the compound mesh...

Link:

pier.unirc.eu/cestino/blade3_work.hdf.gz

Pierpaolo,
a) exclamation mark: I can tomorrow try to load the mesh in FEMAP, it has better tools to investigate what / if there is something wrong.
b) did You figure out what is the difference between 3D-algo '3D-Projection' and 'Hexaeder-meshing'. My opinion up to now, that the HEX-mesher makes pentahedra, if Top- and Bot-face are triangle-meshed, that it makes only hex-elm, if Top- and Bot-faces are purely quadrangel-meshed and that it makes both types of elm if Top- and Bot-faces are meshed with Quadrangle-Preference. Is that wrong or true? And does the 3D-proj-algo the same? Is the only difference that the HEX-algo needs 4 lateral faces, while I see the 3D-proj-algo does even with only 2.

Johannes

I don't know the meaning of the exclamation mark, but the compound mesh is fully OK. I controled it in FEMAP and also in Salome. Moreover it is easy to calculate how many edge- and triangle-elm and nodes will be merged. Building the compound twice (with and without merge) and looking onto the number of nodes and elm, it is clear, that the 50 edges, 204 triangles and 128 nodes of the contact face are merged.

You can even build the compound without merge and use a glue contact in Code-Aster. One advantage is You can use less elements for that for outer part of the wing. The 2 parts must not be coincident in the contacting surface. Look onto the Doc

GlueKont=AFFE_CHAR_MECA(MODELE=MODE,
LIAISON_MAIL=_F(GROUP_MA_MAIT='g_cyl',
GROUP_MA_ESCL='KontFlac',),);

regards, Johannes Ackva

Johannes,

<quote>
You can even build the compound without merge and use a glue contact in Code-Aster. One advantage is You can use less elements for that for outer part of the wing. The 2 parts must not be coincident in the contacting surface. Look onto the Doc

GlueKont=AFFE_CHAR_MECA(MODELE=MODE,
LIAISON_MAIL=_F(GROUP_MA_MAIT='g_cyl',
GROUP_MA_ESCL='KontFlac',),);
<quote>

I think it is still too early for me to undertake such approach. I have been studying the code aster documentation these past few days and that will take long I am afraid.

I also tried and defined a group of surfaces on the top surface of the blade and loaded it with negative pressure. I ran the case and got the post processing images of stresses and displacements. Yet I feel ashamed to own up I still haven't been able to find out in which files the numerical results are stored.... Any hint on this would be HIGHLY appreciated.


Next steps will be getting a D-Box structure out of the solid blade and mesh it. I will use the mesh of the solid blade to get some number out of OpenFoam to use as loads in code aster D-Box structure. So far the only data I have about the loads are those I got from an Octave program of mine which "talks" to xfoil.

Regards,
Pier<br /><br />Post edited by: Pierpaolo, at: 2009/04/04 20:10