[spy_hzdr]

Dear all,

I am using CFX 21.1 and would like to report a problem. When I run a transient simulation that is initialized by a steady-state result, I get the following error:

| An error has occurred in cfx5solve: |
| |
| /trinity/shared/pkg/ansys_inc/v211/CFX/bin/linux-amd64/double-ifo- |
| rt/ccl2flow.exe was interrupted by signal SEGV (11) |
+--------------------------------------------------------------------+


This run of the ANSYS CFX Solver has finished.

More details could be found in the attachement.

I also tried to run the case both with CFX 19.2 and with CFX 21.2. It turns out that only 19.2 works, so the error is likely due to the update of the new version.

It would be nice if any of you could give me a hand.

Best regards,
Pengyu

[ghorrocks]

From your output file I see it does not appear to finish parsing the CCL. This suggests your CCL is the problem. I would read this into CFX-Pre in V21.2 (that is the current version) and output a new def file and try again. This should regenerate the CCL and hopefully correct the problem.

[spy_hzdr]

Originally Posted by ghorrocks

From your output file I see it does not appear to finish parsing the CCL. This suggests your CCL is the problem. I would read this into CFX-Pre in V21.2 (that is the current version) and output a new def file and try again. This should regenerate the CCL and hopefully correct the problem.

Dear ghorrocks,

I reran the case following your suggestion, but the problem is still there (see the *out.text file). The ccl file is as attached and I do not see a problem in it. Any further suggestion? Note that the same case runs well on CFX 19.2.

Best regards,
Pengyu

[ghorrocks]

There is something invalid about your def file. These problems are almost impossible to diagnose, so often it is better to just try again and regenerate it from scratch. So I would make your def file in 21.2 from scratch again. Make sure everything is properly defined, and all your expressions give valid answers (particularly if time=0) - but your error is occurring before the expressions are being evaluated, so I do not think this is the problem.

[spy_hzdr]

Originally Posted by ghorrocks

There is something invalid about your def file. These problems are almost impossible to diagnose, so often it is better to just try again and regenerate it from scratch. So I would make your def file in 21.2 from scratch again. Make sure everything is properly defined, and all your expressions give valid answers (particularly if time=0) - but your error is occurring before the expressions are being evaluated, so I do not think this is the problem.

Dear Glenn,

a short update on the progress.
The case is re-established on V21.2 but the problem is still there. Moreover, this looks like a cluster problem. If I run the case on my local PC it works fine for both CFX 19.2 and CFX 21.2. However, if I run it on a remote cluster the problem would arise if CFX 21.1 or 21.2 is used.

Best regards,
Pengyu

[Gert-Jan]

Impressive ccl. What does it do?

[spy_hzdr]

Originally Posted by Gert-Jan

Impressive ccl. What does it do?

It is used for Euler-Euler simulation of a solid-liquid stirred-tank flow. For details, see https://www.sciencedirect.com/scienc...09250920304073

[ghorrocks]

Yes, there looks like a lot of work in that model

If the model is working when you run locally but fails when you use the cluster then I would do some basic checks eg is the remote cluster using V21.2? Are you using any scheduling software? Is that breaking something?

If those check out OK then I would report it to ANSYS support as it appears to be a bug.