CFD Online Discussion Forums

CFD Online Discussion Forums (
-   ANSYS Meshing & Geometry (
-   -   [ANSYS Meshing] stuck element vs prism layer (

hadrien51 January 12, 2012 14:02

stuck element vs prism layer
2 Attachment(s)
Hello Every body

I would like to mesh a boat with appendices. I would like a fine mesh
with good precision and small elements in the front and the tail of
the boat and in the appendices, plus the prism layers around the boat
and the apendices with different heigths. I use the following strategy:

-first the "surface mesh" is done with curvature,proximity based refinement,
all triangles, octree volume mesh. I delete all 3d elements and I run a
Laplace smoothing

-second I run a Delaunay volume mesh with the existing surface mesh, I
perform some smoothing with quality criterion

-third I create the prism layers all around the boat and appendices,
with "fix marching" toggle on and to finish I perform some smoothing,
freezing prisms and some with all types of elements, always with
quality criterion.

I obtain 25 M of elements , 5M of nodes and the quality is not
good. Of course the prisms decrease the quality, but the worse is
the prism generation produces some elements which occupy the same volume and
stuck elements.

I saw some post here about this problem, so I activated in advanced prism parameters "Prism v10.0" but it doesn't work
"Do checks" and "Do not allow sticking" but it doesn't work

I have always stuck elements.
I tried also without "fix marching direction" but it doesn't work
I tried also without laplace smoothing but it doesn't work


Other problem is the junction between the boat and the apendice with the prism layers.
I have attached a picture

Can anyone help or give me some advice ???

thank you

PSYMN January 13, 2012 10:37

You just need to reduce the transition. The keel is much finer than the hull so it is having trouble with the transition. Refine the mesh on the keel for some distance out from the hull and it will be able to handle the corner... Then have it transition to the larger mesh size over some distance so the prism height transition will be smooth...

hadrien51 January 13, 2012 12:58

Thank a lot Simon !

hadrien51 January 18, 2012 10:38

I reach to obtain 0 stuck elements !!!!!

- help the transition like Simon said
- don't forget to toogle on "do check", "not allowed ..."
-some layers(3) without height info
-split the 3 layers in 10 layers with "fix ratio"
-redistribute 10 layers with fix ratio

if this can help someone !

hadrien51 January 23, 2012 14:00

problem geometry
now, to no waste too much elements, I want to cut the hull to put prisms only on down part and decrease the size element on the top part.
So I produce curves on the intersection of one plan and the hull, I segment the surface hull with the plan (I tried to segment with the curves but it don't work !) , see fig 4.jpg.
I mesh the hull and I obtain this surface mesh (see figure 1,2,3 jpg).
I use the curvature on the down part and not on the top part (see 2.jpg).
Some one can explain me or help me to remove these horrible refinement in the intersection (see 1,3 jpg) and improve the edge (see 2.jpg)

hadrien51 January 23, 2012 14:02

4 Attachment(s)
ooops i forgot the pictures

All times are GMT -4. The time now is 21:33.