Ferrite / Pearlite Formation

solid-solid phase transformations, influence of stresses and strains
Mehnoush
Posts: 28
Joined: Mon Aug 01, 2016 10:31 am
anti_bot: 333

Re: Ferrite / Pearlite Formation

Post by Mehnoush » Thu Feb 14, 2019 9:08 am

Hi again,
please find the *dri file in attachments. :)
Attachments
dt100000_real_mhm.txt
the driving file
(23.29 KiB) Downloaded 146 times

Bernd
Posts: 1222
Joined: Mon Jun 23, 2008 9:29 pm

Re: Ferrite / Pearlite Formation

Post by Bernd » Thu Feb 14, 2019 6:11 pm

Dear Mehnoush,

I had a look on your input file, and I have some remarks. Anyway, what you do is quite complex stuff, and I can only focus on obviously wrong or strange parameters.

- It is very unusual that you change resolution when you restart from your first simulation. Before, Δx was 0.075 µm, after restart it is 0.0075 µm. In principle this is "allowed" (however see below), but the question is how realistic the scenario can be if you just downscale the original microstructure from the fist step by a factor of 10...

- If you want to "reuse" a microstructure in a different context, you should use "restart structure_only" rather than "restart reset_time". I am not sure why the latter works at all for you, given the many parameters which you change between the first and the second step. When using "restart structure_only", you should define realistic initial concentrations for all phases ("input" instead of "equilibrium"), because initial equilibria and all interface linearisations are newly calculated upon restart.

- In any case, you should not change phMin upon restart, because it leads to numerical trouble. You also should not change boundary conditions from isolation to periodic (because the microstructure doesn't fit!).

- I have the impression you should check more often for nucleation (all 0.001s corresponds to 100K interval!).

Bernd

Mehnoush
Posts: 28
Joined: Mon Aug 01, 2016 10:31 am
anti_bot: 333

Re: Ferrite / Pearlite Formation

Post by Mehnoush » Fri Feb 15, 2019 2:43 pm

So i could solve the problem. However, i have a new problem. :? i only have the phase transition in the pearlite phases. although it should happen also in the ferrite. moreover by the increase of the temperature the interfaces becoms very thick and unrealistic.

Bernd
Posts: 1222
Joined: Mon Jun 23, 2008 9:29 pm

Re: Ferrite / Pearlite Formation

Post by Bernd » Mon Feb 18, 2019 2:18 pm

Dear Mehnoush,

please always attach your current version of the driving file, otherwise I don't know how to help you...

Bernd

Mehnoush
Posts: 28
Joined: Mon Aug 01, 2016 10:31 am
anti_bot: 333

Re: Ferrite / Pearlite Formation

Post by Mehnoush » Thu Feb 21, 2019 5:13 pm

Hi Bernd,
so solved the previous problem. so as a summary, we started from a image of the structure with pearlite ferrite grains and by substituting the pearlite grains with austenite we were able to some how reach a equivalent of pearlite ferrite structure (as the MICRESS examples do). Following we heated the obtained structure with 100000 K/s and reached complete austenitization. whoever in the final structure the are still some small unresolved cementite grains. these grains are very problematic for the next step to come, which is the simuation of the cooling during the process. I attached all the *.dri file and the picture hope you can help us futher.
BR,
mehnoush
Attachments
dt100000_real_phas_mcr.png
dt100000_real_phas_mcr.png (16.34 KiB) Viewed 3570 times
dt100000_real_mhm_v4.txt
(23.33 KiB) Downloaded 128 times

Bernd
Posts: 1222
Joined: Mon Jun 23, 2008 9:29 pm

Re: Ferrite / Pearlite Formation

Post by Bernd » Thu Feb 21, 2019 5:55 pm

Dear Mehnoush,

I found two issues in your driving file:

1.) You use a huge interface mobility for the same phase interactions µ(1/1) and µ(2/2) which are about 4-5 orders of magnitude higher than normal. Please keep in mind that there is no "mob_corr" in this case which would reduce them automatically as it is the case for the other interactions!

2.) The interface energies of these same phase interactions σ(1/1) and σ(2/2) are also untypically high (although not to this extent). If the range of interface energies is too high in a triple junction (here between 5.E-6 and 2.E-4 J/cm2, they may get unstable.

Apart from that, please check for error messages which are indicative for specific problems, especially at the beginning. The problem may come from completely other directions which I cannot see without further information...

Bernd

Mehnoush
Posts: 28
Joined: Mon Aug 01, 2016 10:31 am
anti_bot: 333

Re: Ferrite / Pearlite Formation

Post by Mehnoush » Mon Feb 25, 2019 1:04 pm

I will try that. thank you.

Mehnoush
Posts: 28
Joined: Mon Aug 01, 2016 10:31 am
anti_bot: 333

Re: Ferrite / Pearlite Formation

Post by Mehnoush » Tue Apr 09, 2019 1:51 pm

Hi Bernd,
I wanted to know if there is way to reduce the calculation time for bigger input imagesfor example a numerical method. The models work for smaller domain but they take for ever for calculation of bigger images.

BR,
Mehnoush

Bernd
Posts: 1222
Joined: Mon Jun 23, 2008 9:29 pm

Re: Ferrite / Pearlite Formation

Post by Bernd » Tue Apr 09, 2019 2:41 pm

Hi Mehnoush,

The first step here is to find out where all the calculation time goes. Did you already check the .TabP file for that?

Bernd

Mehnoush
Posts: 28
Joined: Mon Aug 01, 2016 10:31 am
anti_bot: 333

Re: Ferrite / Pearlite Formation

Post by Mehnoush » Thu Apr 11, 2019 11:55 am

Hi Bernd,
very nice hint. i am cheking it.
thanks

Post Reply