I'm just starting out with STAR-CD.

When I enter a patch with variable spacing in both the x and y directions and then extrude them, I get cracks near the edges where the cells are thinly spaced.

Why is this and how can I fix it?

Thanks

Jon

before you extrude the patch, try smoothng the vertices using the vsmooth command. for more information type 'help vsmooth' on the command line.

It's an error in checking algorithm. If you are sure that there are no actual cracks there you can just ignore it.

Compress the Cell Delete the unused vertices and compress the vertices The matter would be eliminated

Hello,

I encounter a similiar problem. I built up everything and run the simulations fine. But after I refine the grid, I got cracks.

I have to say, I *did* compress the cells and merged vertexes.

Is this like what Anton Lyaskin said an unreal error message due to the wrong algorithm of checking cells?

Thanks,

JY

Hi,

In your case, it's possible that the "false crack" was due to the refinement algorithm or probably because you didn't choose the create couple option in the refinement command.

I *did* choose the couple option 1 before doubling the mesh. I think it is proper.

Also, I collect the "crack" cells after the cell check and zoom out to see the detail. But I could not find anything strange.

Any idea? (May I just igonre as Anton said?)

Thanks,

Jiaying

My only last idea is the following : select the all entire cells select attached vertices invert the vertices selection delete (+merge +compress)this last selection if not empty After : merge couple and compress it launch the check tools

What did u got ??

Karim

Sorry for the empty post above. I just want to ask, which version do you use? I've had such problems with versions 3.10, 3.10a and early releases of 3.10b. Looks like later the algorithm was corrected.

To my mind the error comes from the nature of crack test - ProSTAR takes a point in the center of every cell face, moves it along the outer normal for some small value "delta" and if it resides in the neighbore cell than there is no crack. In your case the thickness of the cell is less than "delta".

You can try a simple check: "Enlarge" your model (using vscale command) say 1000 times and check it again. If there will be no crack than it's an error in the algorithm and you'll just have to set the geometry scale factor 0.001 before you'll run the calculation.

Regards, Anton