CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   ANSYS Meshing & Geometry (https://www.cfd-online.com/Forums/ansys-meshing/)
-   -   [ICEM] ICEM CFD hex mesh with very small y+ (https://www.cfd-online.com/Forums/ansys-meshing/127249-icem-cfd-hex-mesh-very-small-y.html)

Shenren_CN December 6, 2013 16:05

ICEM CFD hex mesh with very small y+
 
I try to generate the all-hex mesh and have had the blocking done properly.

The mesh quality is very good is I set the edge to be uniformly spaced.

However, whenever I try to refine the boundary layer mesh with small y+,
by setting the edge pre-mesh params to a small length on one end and
use geometric growth for the edge, and then COPY THE PARAMETERS
TO ALL PARALLEL EDGES, the resulting mesh always have negative vol
cells.

I suspect it is because that the PARALLEL EDGES all have (slightly) different
length, and thus I don't know what really happens when the PARAMETER
for one edge is copied to the rest.

Has anyone experienced the same issue, and any suggestions on how to
resolve this?

Cheers,
Shenren

khoopes December 6, 2013 17:10

I have had this problem before where it will break for small values of first element distance. I think the problem is that you are trying to make elements that are smaller than the underlying triangulation of the geometry that ICEM defaultly uses to place the elements. You can either adjust this triangulation to make it finer or use the project to b-splines option and the projection limit to not use the triangulation for node placement.

adjust triangulation
Settings -> Models/Units -> Triangulation
make it really low

or (the first one is kind of a quick fix)

Project to Bsplines
Settings -> Meshing Options -> Hexa Meshing
turn it on

You will have to adjust the projection limit as well, from the help:

"The Projection limit is set to a non-zero value in cases where you want to keep the nodes on the edges and avoid projection to the underlying surfaces. This option is typically used for Navier-Stokes grids where the grid spacing is small relative to the geometry tolerance. Allowing nodes within a gap in the geometry to project would skew the elements. With a value P set to slightly larger than the gap, the nodes would instead be interpolated. The value may have to be set by trial and error depending on skewness or negative determinants being reported by pre-mesh quality checks."

Shenren_CN December 6, 2013 17:31

Thanks for the reply, Kevin. I'll give a try and let you know. Indeed, my first element distance to wall is extremely small in order to get y+ around one.

BTW, Did you happen to have encountered the issue of not having all the edges of the same length when "COPY PARAMETERS TO PARALLEL EDGES". I originally thought this has caused my problem.

Cheers,
Shenren

Quote:

Originally Posted by khoopes (Post 465185)
I have had this problem before where it will break for small values of first element distance. I think the problem is that you are trying to make elements that are smaller than the underlying triangulation of the geometry that ICEM defaultly uses to place the elements. You can either adjust this triangulation to make it finer or use the project to b-splines option and the projection limit to not use the triangulation for node placement.

adjust triangulation
Settings -> Models/Units -> Triangulation
make it really low

or (the first one is kind of a quick fix)

Project to Bsplines
Settings -> Meshing Options -> Hexa Meshing
turn it on

You will have to adjust the projection limit as well, from the help:

"The Projection limit is set to a non-zero value in cases where you want to keep the nodes on the edges and avoid projection to the underlying surfaces. This option is typically used for Navier-Stokes grids where the grid spacing is small relative to the geometry tolerance. Allowing nodes within a gap in the geometry to project would skew the elements. With a value P set to slightly larger than the gap, the nodes would instead be interpolated. The value may have to be set by trial and error depending on skewness or negative determinants being reported by pre-mesh quality checks."


khoopes December 9, 2013 08:50

I have never had an issue with different sized parallel edges. That has always worked fine for me.

anaekh July 25, 2017 23:11

Thanks for this post. It helped me in solving my problem

OVS July 31, 2017 15:07

Amazing post. I've been looking for this answer for months...

Thanks!


All times are GMT -4. The time now is 11:57.