I-Radiation Equation problem with CFX Parellel
 

All equations' information about the first iteration appear amostly at the same time in cfx output file when submit a little case on a local machine, but if using cfx parellel process we can find that computers spend a long time on calculating I-Radiation equation of the first iteration and it will be ok at the rest iterations.Do anybody know how to solve that? ......I have wait for the finish of the first iteration for almost one day,but I-Radiation doesn't appear yet.It make me ill..

Re: I-Radiation Equation problem with CFX Parellel
 

Dear Scott,

Are you using the Discrete Transfer, or Monte Carlo models? For the DTM and MC models, the first iteration is a lot more expensive than any. For example,

DTM: ----

1 - Create internal topology information for ray tracing.

2 - Creates the track file, and stores the tracks to disk.

3 - After the first iteration, (1) and (2) are no longer needed. Therefore, faster turn around

MC: ---

1 - Still creates topology information for ray tracing.

2 - No need for (1) after first iteration.

In 10.0, the algorithm for crossing domain interfaces is very slow (DTM and MC) if the domain interfaces have a large number of element faces. In 11.0, this problem has been resolved (contact local support for the latest 11.0 builds).

Beyond that, no idea what the problem could be.

Opaque

Re: I-Radiation Equation problem with CFX Parellel
 

Dear Opaque, Thanks for your reply,I used Monte Carlo models for thermal radiation simulation,and according to your reply,i think what i can do is to contact support for new release.Thank you.