MICRESS Forum

dear Bernd,
I want to simulate the interphase precipitation during the austenite → ferrite transformation, I built a driver file for Fe-Mn-Al-C-Ni alloy, where the second phase precipitation is set to kappa phase, and make It is precipitated on the grain boundaries of austenite and ferrite, but the results are very unsatisfactory, the following is how I create a driver file and hope to get your help.
shoajie lyu

Dear shaojielv,

I am wondering why you set phase 2 to stoichiometric for Al and C. At the same time you assume the nple model for Al in phase 2 - this cannot work!
Generally, you need to check whether nple can be used at all for Al, because the model assumes full solubility in the whole composition range in both phases.

Bernd

Hi Bernd
First of all, thank you for your reply. I'm sorry that I don't quite understand the stoichiometry in this software, so there is a mistake in setting. Do you have any suggestions? Or how do I optimize my model?

Dear shaojielv,

I have no experience with this type of alloy. In which way the results are unsatisfactory? Do they improve when you remove the stoichiometric condition and the "nple" definition for Al?

Hi Bernd,

First, there was a problem with the simulation,
//
Warning: Demixing in interface FCC_A1/BCC_A2, component MN
Warning: Demixing in interface FCC_A1/BCC_A2, component NI
Updating of diffusion data from database...
//
The second is that evolution is very slow, running only 20 seconds a night.
What do you think caused this?

Dear Bernd,
And why doesn't ferrite evolve during the phase transition? I checked the program again and adjusted the relevant parameters, but there was still such a phenomenon. I was curious why. I hope you can help me. Thank you very much.

Dear shaojielv,

Based on your "Demixing" messages, I suggest to use diagonal redistribution, which is often needed in case of high-alloyed systems with strong interactions between the alloyed elements. You can do that by putting the keyword "diagonal" in the input section for "Numerical Parameters" of the "Concentration Solver" (just where you had before the stoichiometric definition). This may alter the behavior you observe. In any case, the next step should be to check the data of the initial equilibria, which are written to the .log file upon the first nucleation check.

With respect to performance, you should have a look on the .TabP output file (see here).

Bernd