I've been struggling with NX 10 and the ESC solver for some time now and unfortunately was never able to achieve any robust sim model.
Usually there's some point where everything runs fine but after a small change (e.g. changing emissivity or thermal load slightly) it crashes with some ominous "synchronization error", "No temperature BCs specified for steady state run" (which btw is nonsense) or the like. This happens sometimes before the actual solver even starts and sometimes in the middle of the solving process while convergence is looking promising.
And to make things worse it also happens without actually changing anything. Just starting a previously successful solving process over again will make NX crash with a likelyhood of roughly 33%.
Now there's a warning in the .log and MSGF file that keeps haunting me from the beginning (and which I was told is just one of the numerous false-warnings in the log files). It kind of looks like this:
** WARNING **
** File C:\Users\xxx\Documents\NX\thermosim\Strahlungsuntersuchung\wu
** cannot be deleted for some reason in routine opencp,
** possibly because it is currently being accessed.
BTW: There's no such file as "wu" - I guess it's been cut off. The MSGF file warns a lot about flow.tmp and graph.tmp files which it isn't able to delete.
However, sometimes the crash happens one iteration after such a warning and it all ends with:
Synchronization error: flow solver has crashed or is not responding.
So I think it might be related to my problem.
After a crash I've often tried deleting all result-files and even the mesh-files manually (which works just fine) and sometimes NX manages to deliver a result after that - but sometimes it doesn't. I think it doesn't affect the probability of a successful simulation.
Just a few facts that might be worth mentioning:
) I use NX outside the TC context. So all files are stored locally on a Win7 installation.
) I have - more or less - admin permissions.
) There's no space or any odd characters in the path this all happens (as you can see above).
) The workstation I'm using is capable of running such simulations; it's got plenty of RAM and there's hundreds of GB space left on the HDD.
It'd be great if someone could help me on that or give me some hints on where to have a closer look. Don't hesitate to ask me if there's some files I can provide.