Error during simulation

Discussion of numerical issues which are not directly linked to specific applications
betleenkim
Posts: 41
Joined: Tue May 07, 2013 1:01 pm
anti_bot: 333

Re: Error during simulation

Post by betleenkim » Wed Jun 05, 2013 1:59 am

Dear bernd

Thank you for the file and your reply.
I will report you the result.

In addition, does this error cause rebooting of the system?
I found that the simlation is stopped with error message in the window and sometimes my computer automatically reboots during simulation.

Best regard

Bernd
Posts: 1138
Joined: Mon Jun 23, 2008 9:29 pm

Re: Error during simulation

Post by Bernd » Wed Jun 05, 2013 5:04 pm

Dear Seawoong,

I cannot see any relation between MICRESS and rebooting of your system!

This sounds more as if your system configuration is unstable...

betleenkim
Posts: 41
Joined: Tue May 07, 2013 1:01 pm
anti_bot: 333

Re: Error during simulation

Post by betleenkim » Fri Jun 07, 2013 4:20 am

Dear Bernd

I am happy with the Beta version (6.006) of MICRESS. I can do simulation without any error.
However, the result from MICRESS 6.006 version and MICRESS 6.006 version (Beta) is quite different as below.
Different result between 6.003 and 6.006(beta).jpg
Different result between 6.003 and 6.006(beta).jpg (143.02 KiB) Viewed 2133 times
I used same database (TCFE7 + MOBFE2) and same interface mobility.
Where does the difference come from? Is it due to the beta version of MICRESS?
In addition, it seems that the simulation speed is also quite slower than the 6.003 version even though i do simulation in much faster PC (CPU frequency : 4.2GHz).

Best regard.
Attachments
Different result between 6.003 and 6.006(beta).jpg
Different result between 6.003 and 6.006(beta).jpg (143.02 KiB) Viewed 2133 times

Bernd
Posts: 1138
Joined: Mon Jun 23, 2008 9:29 pm

Re: Error during simulation

Post by Bernd » Fri Jun 07, 2013 8:18 pm

Hi Seawoong,

that is very strange :shock:

It seems that the solid-solid interaction is not doing anything with the new version. I cannot tell you anything more at the moment, please send me the input files, and I will see why it behaves differently in this case...

Bernd

betleenkim
Posts: 41
Joined: Tue May 07, 2013 1:01 pm
anti_bot: 333

Re: Error during simulation

Post by betleenkim » Sat Jun 08, 2013 6:33 am

Dear Bernd

Thank you for your reply.

I sent the input files via your e-mail.

Best regard

Seawoong Lee

Bernd
Posts: 1138
Joined: Mon Jun 23, 2008 9:29 pm

Re: Error during simulation

Post by Bernd » Tue Jun 11, 2013 11:56 pm

Dear Seawoong,

I tried but cannot reproduce your results - phase 2 is not stable at all :shock: !

Are you sure that you sent me the right driving file? I have seen that the one you sent me uses restart - did you obtain the different results using restart?
Furthermore, I used the .GES5 file which you sent me last time - is it correct?

Bernd

betleenkim
Posts: 41
Joined: Tue May 07, 2013 1:01 pm
anti_bot: 333

Re: Error during simulation

Post by betleenkim » Wed Jun 12, 2013 2:58 am

Dear Bernd

Thank you for your reply.
I sent you the driving files and the result files simulating with MICRESS 6.006 version.
Please check the files.

In addition, could you let me know once you get the e-mail?
I am little bit frustrated these days since i have not gotten any simulation result since i have left Germany.
So i really want to solve these problems as soon as possible.

Best regard

Bernd
Posts: 1138
Joined: Mon Jun 23, 2008 9:29 pm

Re: Error during simulation

Post by Bernd » Thu Jun 13, 2013 5:07 pm

Dear Seawoong,

no reason to to frustrated!

I think it does not make so much sense to really trace down why Version 6.003 and 6.006 seemingly are behaving different. It is more important to bring you back on track! And, as I said in my last post, the simulation is very instable at present (what probably also could explain the difference of the versions...).

The main problem in the simulation is the high Al content which makes the thermodynamic system behaving not so nice. You already set Al stoichiometric in BCC. But there is also a pseudo-binary demixing of Al in the 0/2 interface, and it is impossible to set Al also stoichiometric in FCC (otherwise it would be impossible to simulate the 1/2 interaction).

Therefore, my proposal is to use the new multi-ternary extrapolation scheme which will be officially available in MICRESS 6.1. Although an experimental version of this feature was already available in the 6.006 version, I will send you a new executable by mail.

To avoid the actual numerical problems with FCC, you should define a ternary subsystem Fe-Mn-Al by defining

interaction 2 3

instead of the stoichiometric condition.

Another issue which you will have to change is that many of your seed type definitions are not reasonable or even not allowed (the new version will complain...): If you define phase 1 as reference phase and as phase of the new seed, this is not allowed if no phase interaction is defined between phases 1/1! It is also impossible to calculate any driving force for this type of nucleation, it would make sense only in case of recrystallisation...
I would recommend you to reduce the number of seed types to the minimum which you really need!

Some further comments:

- There are two initial seeds, but one is outside the domain. Due to the periodic boundary conditions on the east and west side, two half dendrites are formed which have the same grain number, and not two different dendrites. You are aware of that?
- The interface energy of the 1/2 interface is too high in my opinion, to my knowledge it should be similar to 0/1 and 0/2.
- If "fd_correction" is used (as recommended), the definition of the interface thickness is slightly different: An interface thickness of 4 cells corresponds to 5 cells without fd_correction.
- try to reduce the size of the domain while you are optimizing numerical parameters! The simulation must be fast to be able to try out different combinations of parameters.
- there is potential for performance improvement. We should discuss that at a later time.

Please tell us about your success!

Bernd

betleenkim
Posts: 41
Joined: Tue May 07, 2013 1:01 pm
anti_bot: 333

Re: Error during simulation

Post by betleenkim » Fri Jun 14, 2013 3:34 am

Dear Bernd

Thank you for the e-mail. :)
Now i am doing a simulation with MICRESS 6.010 version.
I hope that i can get good result.

And also thank you for the advice in your last post.
I will follow them to achieve reasonable result.

I will inform you what is going with the new version.

Best regard

betleenkim
Posts: 41
Joined: Tue May 07, 2013 1:01 pm
anti_bot: 333

Re: Error during simulation

Post by betleenkim » Wed Jun 19, 2013 10:01 am

Dear Bernd

The simulation is going well so far.
Thank you for the supports. :D

Best regard.

Post Reply