Arbitrary Matching & Partial Boundaries
 

I call upon the forum for some well-needed advice with regards to partial boundaries in Star-CD 3.1.

I'm trying to match a semi-circular surface with a larger rectangular surface for the purpose of modeling an air handling unit. At the interface, the cells are approximately 1" cubes on either surface (except for the outer portion of the semi-circle, where the aspect ratio is slightly larger).

The problem is as follows:

I apply partial boundaries before implementing the arbitrary match command, yet I have cracks in the model at the partial boundary interfaces (indicating to me that the cells are not matched). Then, when I apply partial boundaries again in order to get these specific cells to match, my crack problem is solved, but there is no more partial boundary (i.e. there are holes at the interface). Any suggestions?

Kindly help a Star-CD rookie in need...

Re: Arbitrary Matching & Partial Boundaries
 

Why don't you try to "chisle" the imprint of the semi-circular shape on the rectangular surface? By "chisle" I mean cut some of the hexahedral elements into two prismatic elements. You can then make some integral matches one cell layer deep into the rectangular domain, to take care of the prisms. (i.e. Cpmatch the two prisms to one hex)

Once you have "chisled" the imprint, put a spline on the semi-circle and then move the points from your newly "chisled" perimeter onto the spline. Take a look at the following commands (VSPM, VSPF and VSPC).

Now you will have a common perimeter arbitrary interface, and no partial boundaries.

Your other option is to try and carry the semi- circular topology into the rectangular domain (i.e. point to point connectivity at the interface).

I would stay away from partial boundaries.

Have Fun!

Chinor

Re: Arbitrary Matching & Partial Boundaries
 

It should be "chisel".

Never was a good speller.

Re: Arbitrary Matching & Partial Boundaries
 

Thanks, Chinor. I'll use your advice.

G

Re: Arbitrary Matching & Partial Boundaries
 

Hello,

I would to know how do you split a hexaedral element into two prismatic element with Star-CD in order to chisel your mesh.

Thanks.

HV

Re: Arbitrary Matching & Partial Boundaries
 

You can use the "cdx samm 1" to create a degenerated samm cell of type 1 which has an additional diagonal edge on one face.

Re: Arbitrary Matching & Partial Boundaries
 

Thank you for your answer.

HV