problem on restart the simulation

[swh2011]

Dear Bernd:
I am simulating the nucleation, growth, coarsening of precipitates in matrix phase using Micress 5.501 plus TQ coupling, and there is a problem. The nucleation, growth , early coarsening process are ok. But in later coarsening process, when I restart the simulation, the .log file shows:
Time t = 0.97000E+07 s
Fraction of phase 0: 0.00000
Fraction of phase 1: 0.83947
Fraction of phase 2: 0.16053

**********************************************
* Begining of simulation *
**********************************************
complete relinearisation!
complete relinearisation!
complete relinearisation!
Error number 3 in Interface 1 30
time: 9700268.0000000000 s
ERROR RETURN FROM solveCNewton BECAUSE THERE HAVE BEEN
Error number 3 in Interface 1 44
time: 9700293.0000000000 s
ERROR RETURN FROM solveCNewton BECAUSE THERE HAVE BEEN
complete relinearisation!
Error number 3 in Interface 1 8
time: 9700346.0000000000 s
ERROR RETURN FROM solveCNewton BECAUSE THERE HAVE BEEN
Error number 3 in Interface 1 39
time: 9700362.0000000000 s
ERROR RETURN FROM solveCNewton BECAUSE THERE HAVE BEEN
complete relinearisation!
Error number 3 in Interface 1 2
time: 9700412.0000000000 s
ERROR RETURN FROM solveCNewton BECAUSE THERE HAVE BEEN

Can you help me to solve this problem?
Thank you very much
swh

[Bernd]

Dear swh,

this is very little information to find out what is happening. The error message is cut, but it means that the number of iterations for finding a valid quasi-equilibrium condition is exceeded. This information as such is not really helpful (PS: in version 6.003 the error messaging for TQ coupling is completely reworked!), it just shows that MICRESS is not working in a sound way. On the other side, there is still no hint that the problems are severe!

For a more explicite answer I would need more information:
- which is the system you are simulating (elements, phases, database)
- is MICRESS recovering later, or does it continue with much more of these messages?
- what is seen in the screen output? In case of numerical instabilities, the screen output (or the _TQ.txt output with Windows) gives much more information.
- does it only occur with restart, or also if you continue without restart?
- why did you restart/what did you change in the driving-file for restart?

If possible, you should also attach the .dri file for which the problem occurred.

Bernd

[swh2011]

hi, I send you the files to your email.
Thanks
swh

[Bernd]

Hi swh,

got the files, did not know that you are doing stress-coupling...

I have to admit that I am not a specialist for stress calculation, so I cannot really see wether this part is doing well. But at least it could explain why the problem is ocurring at a such late stage in ripening, if stress is accumulated somehow.

To know more about the relinearisation problem, I would need the .TabLin output...

Thanks!

Bernd