Recrystallization

ripening phenomena, dislocations, grainboundary topology
shaojielv
Posts: 100
Joined: Thu Apr 01, 2021 2:16 pm
anti_bot: 333

Recrystallization

Post by shaojielv » Tue May 17, 2022 4:34 pm

Dear Bernd,

I would like to ask some questions about recrystallization. I coupled the thermodynamic database through TQ-interface, and coupled the phase field and concentration, and used the mean_dislocation model, but the simulation shows that the element concentration does not affect the recrystallization results. Do you think this is What is the reason? Looking forward to your reply.

Lv

Bernd
Posts: 1504
Joined: Mon Jun 23, 2008 9:29 pm

Re: Recrystallization

Post by Bernd » Tue May 17, 2022 4:57 pm

Dear lv,

As long as you apply the recrystallisation model to a single-phase system (which is the typical application case), there is no concentration dependent term which would influence the reX-process. This is because in same-phase interactions, there is no effect of concentration on the driving force for grain growth. Indirect effects of concentration e.g. on the interface mobility and energy or on the properties of the dislocations are not automatically taken into account and must be given explicitly by the user.
However, if you would apply the reX-model to multiphase problems, there would be a dependency because the effect of the reX-energy (or dislocation density) on the driving force would add up to the chemical contribution.

Bernd

shaojielv
Posts: 100
Joined: Thu Apr 01, 2021 2:16 pm
anti_bot: 333

Re: Recrystallization

Post by shaojielv » Fri May 20, 2022 2:21 pm

Dear Bernd,

First of all, thank you for your letter. I understand what you mean. In this case, I have a question about polyphase that I would like to ask you.
Based on your answer, I started to simulate the austenite-ferrite transformation, and wanted to recrystallize the ferrite. Regarding the ferrite recrystallization model, the maximum recrystallization energy occurs only when the phase and nucleation are defined. The options for the value and the minimum value, of course, this is certainly no problem, but why does the result of my simulation not recrystallize? Can you share your experience with me? Thank you very much and look forward to your reply.

Lv

Bernd
Posts: 1504
Joined: Mon Jun 23, 2008 9:29 pm

Re: Recrystallization

Post by Bernd » Fri May 20, 2022 5:19 pm

Dear lv,

Can you please further specify your question? You are interested in the effect of stored energy on the austenite-ferrite transformation? Or just in "normal" recrystallisation of ferrite?

Bernd

shaojielv
Posts: 100
Joined: Thu Apr 01, 2021 2:16 pm
anti_bot: 333

Re: Recrystallization

Post by shaojielv » Fri May 20, 2022 5:50 pm

Dear Bernd,

Yes, I am interested in the austenite-ferrite transformation of stored energy, and I also want to do ferrite recrystallization.
I will further explain my problem, I am doing austenite-ferrite transformation and recrystallizing ferrite, so when setting up the model I only recrystallize ferrite, austenite does not Recrystallization occurs. In this case, my initial microstructure phase austenite does not involve stored energy, so only ferrite needs to set the recrystallization model and energy threshold. In this case, the ferrite grains have no recrystallization energy. When nucleating, only the maximum and minimum recrystallization energies need to be assigned, but in this case, my ferrite does not recrystallize. I don't know why, I don't know if I have expressed it clearly, if not, I can send the driver file.

Lv

Bernd
Posts: 1504
Joined: Mon Jun 23, 2008 9:29 pm

Re: Recrystallization

Post by Bernd » Fri May 20, 2022 6:43 pm

Dear lv,

Now I got what is your problem: Because ferrite is not present at the beginning of the simulation, it is not possible to assign a stored energy or dislocation density. From a physical point of view it is correct that freshly formed ferrite has no stored energy, and deformation typically is induced in reality afterwards by thermal thrinkage or rolling.

As explaind in this thread, MICRESS is not yet equipped for describing such dynamic recrystallisation cases. But there are two ways of working around:
One is to assign a recrystallisation energy for new nuclei of ferrite (although this may seem physical nonsense).
The other one is to use an extra pseudo-nucleation path to indirectly assign stored energies at a later stage by a virtual "deformation" process. This method is using "add_to_grain" in such a way that no new grains are formed, but the properties of the already existing grains are changed. To do so, you would use the sub-option "parent_grain" and nucleate new ferrite seeds on top of the existing ferrite. Those seeds would not be set (because the same grain already exists there), but the properties (including reX-energy) of the existing ferrite would be altered.

Is that what you intend?

Bernd

shaojielv
Posts: 100
Joined: Thu Apr 01, 2021 2:16 pm
anti_bot: 333

Re: Recrystallization

Post by shaojielv » Sat May 21, 2022 3:11 pm

Dear Bernd,

Yes, that's right, as you said, that's my intention, the first one you mentioned assigning recrystallization energy to ferrite, I think this is a good practice for users, please this How should it be done? Can you talk more about that?
I carried out the relevant simulation according to the second method you mentioned. When my ferrite nucleated at the austenite grain boundary, I added add_a_grain and selected parent_grain. The result of this operation is: only one of them Austenite becomes ferrite, and the transformation takes place only in this grain (result shown in Figure 1); when ferrite nucleates within the grain, I added add_a_grain, and also selected parent_grain, the result At the nucleation temperature I set there was a direct transformation (mutation) of austenite to ferrite, and over time the ferrite stopped growing (results shown in Figure 2). I don't understand why this is, I attached the picture, I hope you can provide me with some reference learning advice, thank you very much.

Lv
Attachments
2.png
2.png (144.99 KiB) Viewed 4810 times
1.png
1.png (20.87 KiB) Viewed 4810 times

Bernd
Posts: 1504
Joined: Mon Jun 23, 2008 9:29 pm

Re: Recrystallization

Post by Bernd » Mon May 23, 2022 10:53 pm

Dear lv,

With the first method, if you have set ferrite as recrystallization phase, you are required to give a range of reX-energies when you want to nucleate ferrite at austenite grain boundaries. You should chose them bigger as the threshold value, so that when they start growing no recrystallisation will happen. Afterwards you nucleate ferrite on ferrite for triggering recrystallisation of ferrite. Here, you should chose a reX-energy of zero for the nuclei (or something less than the threshold value).

For the second method, you need to take care, when using add_to_grain, that ferrite is the parent phase. Otherwise you switch the austenite grains to ferrite! To achieve this, you should set ferrite as first substrate phase and as reference phase, while austenite should be second substrate phase. The second substrate phase just makes sure that nucleation is allowed at ferrite/austenite interfaces, but it is not chosen as parent phase.

Bernd

shaojielv
Posts: 100
Joined: Thu Apr 01, 2021 2:16 pm
anti_bot: 333

Re: Recrystallization

Post by shaojielv » Tue May 24, 2022 10:52 am

Dear Bernd,

First of all, I am very happy to receive your reply. In your reply, I also understood the establishment of the ferrite recrystallization model.
Based on your explanation, I'm going to start the simulation from the second method, and use ferrite as the first matrix phase and reference phase, and austenite as the second matrix phase. And the model is established, the following is my driver file, can you see if there is any problem with the model?

lv
Attachments
ReX_FeCMn.dri
(24.36 KiB) Downloaded 150 times

Bernd
Posts: 1504
Joined: Mon Jun 23, 2008 9:29 pm

Re: Recrystallization

Post by Bernd » Wed May 25, 2022 8:52 pm

Dear lv,

Nucleation is not correct as you do it:

1.) Before using method 2 with "add_to_grain" give ferrite a reX-energy, you need to create ferrite with a first "normal" seed type (which is missing)

2.) In the second seed type (the one with "add_to_grain" which you have right now) the phase of the new grains must be 2 (ferrite), otherwise you will create austenite instead of changing the properties of ferrite!

3.) You should not define "recrystall" for austenite in the phase data input if you do not want to recrystallize austenite

4.) As it is the goal of method 2 to put reX-energy into the ferrite grains, you should not define the range of the reX-energy from 0 to 0 in the nucleation data

5.) When nucleating ferrite on ferrite in seed type 2, you will not have a chemical driving force. Thus, you should define a critical undercooling <0 to enable nucleation

Bernd

Post Reply