CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Pointwise & Gridgen (http://www.cfd-online.com/Forums/pointwise/)
-   -   Block initiation fails, Pointwise (http://www.cfd-online.com/Forums/pointwise/82148-block-initiation-fails-pointwise.html)

filipwa November 18, 2010 06:06

Block initiation fails, Pointwise
 
I need to do calculations on a plain wing and i am trying to do the mesh in pointwise. I chose to put the wing inside a cube, which will represent the far field, with the dimensions 20x20x20 m. The wing itself is 10 m long.

After building the surface mesh of the wing i built a block between the wing and the 6 domains representing the far field (cube). To do this i used the command "Create, Assemble Special, Block". Which seems to work without any problems.

However, when i try to initiate the block, half way through the process, i get an error message saying "There are intersecting triangle edges".

Why is this? And how should i go about correcting it?

Any suggestions would be very much appreciated!

rmatus November 19, 2010 09:56

Causes of intersecting triangle edges
 
There are several things that could cause intersecting triangle edges:

1. There could be overlapping or intersecting surface grids. This is not too likely, since you were able to assemble a block, but it does not hurt too visually check for this when you are assembling the block.

2. If you have used any quadrilateral surface grids in the face of a tetrahedral block, Pointwise will automatically put pyramids on top of these to provide point-to-point matching between the quads and tetrahedra. It is possible for these pyramids to intersect with each other, particularly if they are on high aspect ratio quads near concave corners. You can visually check for this by selecting the block and going to the Grid, Solve panel. This will display the pyramids. You can adjust pyramid heights in the Attributes, Pyramid pane if necessary.

3. Highly skewed triangles on the surface grids could also cause this message since they could cause odd tetrahedra to be generated in the volume. It is a good idea to check the surface mesh quality before generating a volume mesh. You can use your quality criteria of choice. I usually try to get all the equ-iarea skews below 0.9.

If none of these help, could you post some pictures of your grid or the actual surface grid files? To get more specific in diagnosing this problem, I would need more detail.

Thanks,
Rick

filipwa November 23, 2010 23:51

Thanks for your reply! Keeping the equiarea skewness to below 0.9 helped me in this case! However, i ran into a new problem though...

I was trying to redo the whole thing and instead looking at only half the wing utilizing a symmetry plane.. Again the block initiation crashes half way through the process saying "There are intersecting edges and triangles".. (The equiarea skewness is kept below 0.9.)

Have been trying the make the mesh finer but it doesnt help. How can I see where those intersecting edges and triangles are?

(Rick, I would be happy to send you the file! Email?)

filipwa November 24, 2010 00:42

Tried to upload the file here but it was too large.

siri November 24, 2010 02:36

Did you try to control the mesh skewness further at the interface of wing & plane of symmetry; see if it helps to reduce skewness criteria to less than 0.65 ( i once used a max limit of 0.7 for meshing of airfoil sections in Fluent). does pointwise have other quality metrics like aspect ratio etc

filipwa November 24, 2010 05:36

Quote:

Originally Posted by siri (Post 284546)
Did you try to control the mesh skewness further at the interface of wing & plane of symmetry; see if it helps to reduce skewness criteria to less than 0.65 ( i once used a max limit of 0.7 for meshing of airfoil sections in Fluent). does pointwise have other quality metrics like aspect ratio etc

yes, the aspect ratio can be controlled.. what would you recommend as a highest value?

rmatus November 24, 2010 10:11

What type of surface mesh?
 
Is your surface mesh all triangles or does it have some quads in it?

I PM'ed my email address to you in case you want to send me the files.


All times are GMT -4. The time now is 07:11.