Unmatched boundary
 

The CFD family,

I am have read somewhere in the manual that if I have unmatched boundary then I should not submit my work to VOLMSH with the the REBLOCK toggled on, in the Analysis of CFXBuild4.4. However, I have done the contrary and yet my simulation results compare quite welll with experimental data. Why is this so?? I want be certain that my results are not just a fluke!

Any idea please!

Regards nyatoto

Re: Unmatched boundary
 

Hi Nyatoto,

The reason is after equivalencing the nodes, if any elements match up on either side of the unmatched interface the reblocker will think it is a block, and hence will construct a block which crosses the unmatched grid interface.

If this happens, the solver will bomb out with an unmatched grid error. If you didn't get this error (and it sounds like you didn't), then there should be no problem with your simulation.

Glenn

Re: Unmatched boundary
 

Hi Glenn,

Thanks for you response. What you say really makes sense.

I did receive a warning when VOLMSH was creating the geo file and it says, "UNMATCHED BOUNDARY NAMES WILL CHANGE". I ignored this warning.

Now here are three other things I do not understand:

1. When I read the geo file in the RUN SETUP I did notice this: PATCHES: <INTER BLOCK BOUNDARY PATCH NOT DISPLAYED>

Now I ask not displayed but is it there??

2. When I browsed the OUT PUT FLILE I saw that the names of the unmatched boundary were just as I had specified in the BUILD.

3.Looking at the results in the POST PROCESS (ANALYSE), I saw that the unmatched boundaries appear as I had specified.

So the unmatched boundary names appear everywhere except in the geo file, and VOLMSH tells me that they will be renamed!!!

regards Nyatoto