The desired effect is to form a ferrite at the triple junction, and then to generate lamellar beads at the grain boundary.
According to the iron carbon phase diagram analysis, the difference between the formation temperature of ferrite and pearite is about 70℃, and the cooling speed is 0.3℃ / s.
But in our model, ferrorites always grow too fast, resulting in the entire simulation domain before the pearlet is nucleated.
We tried to reduce the interface energy between the phases, but the organization effect was very bad.
Want to ask that we can change those parameters to solve the above problems.
solid-solid phase transformation
solid-solid phase transformation
Last edited by kaiyang on Tue Oct 31, 2023 2:41 am, edited 1 time in total.
Re: solid-solid phase transformation
Dear Kaiyang,
What do you mean with "organization effect"?
I did not find any obvious problems in your driving file. There may be different numerical reasons why you do not get cementite as expected. Depending on whether you do observe nucleation events for cementite or not, I could imagine the following reasons:
a) no nucleation of cementite happens at all (see .log-ouput)
- you have specified a too high critical nucleation undercooling for cementite which is not reached
- growth of ferrite is unstable, so that nucleation is not possible at 1/2-interface
- missing updating of thermodynamic data. Please apply updating of linearisation data either for all interfaces (in "database" section), or for each phase interaction separately (in phase interaction data).
- something went wrong when creating the .GES5-file
b) nucleation happens, but cementite cannot grow
- limited number of nucleation events of cementite (500) is spent before growth conditions are suitable
- growth of cementite is unstable, e.g. because of too high interface energy or interface mobility
- cementite seed is overgrown by ferrite before it can get bigger
Bernd
What do you mean with "organization effect"?
I did not find any obvious problems in your driving file. There may be different numerical reasons why you do not get cementite as expected. Depending on whether you do observe nucleation events for cementite or not, I could imagine the following reasons:
a) no nucleation of cementite happens at all (see .log-ouput)
- you have specified a too high critical nucleation undercooling for cementite which is not reached
- growth of ferrite is unstable, so that nucleation is not possible at 1/2-interface
- missing updating of thermodynamic data. Please apply updating of linearisation data either for all interfaces (in "database" section), or for each phase interaction separately (in phase interaction data).
- something went wrong when creating the .GES5-file
b) nucleation happens, but cementite cannot grow
- limited number of nucleation events of cementite (500) is spent before growth conditions are suitable
- growth of cementite is unstable, e.g. because of too high interface energy or interface mobility
- cementite seed is overgrown by ferrite before it can get bigger
Bernd
Re: solid-solid phase transformation
Dear Bernd,
When I was doing a case related to Fe-C solid state phase transition, when I coupled the Mn element in the model, the following error appeared in the run log. I hope to get your help.
trying hard phases 2 3 level: 7 zp= 65758 error= 3
trying hard phases 2 3 level: 8 zp= 163858 error= 3
trying hard phases 2 3 level: 5 zp= 74298 error= 3
trying hard phases 2 3 level: 3 zp= 74298 error= 3
trying hard phases 2 3 level: 8 zp= 53752 error= 3
trying hard phases 2 3 level: 7 zp= 63800 error= 3
trying hard phases 2 3 level: 7 zp= 205031 error= 3
trying hard phases 2 3 level: 7 zp= 163762 error= 3
trying hard phases 2 3 level: 4 zp= 163762 error= 3
trying hard phases 2 3 level: 9 zp= 58974 error= 3
trying hard phases 2 3 level: 9 zp= 30391 error= 3
trying hard phases 2 3 level: 6 zp= 30391 error= 3
trying hard phases 2 3 level: 9 zp= 152717 error= 3
trying hard phases 2 3 level: 8 zp= 738 error= 3
trying hard phases 2 3 level: 8 zp= 738 error= 3
When I was doing a case related to Fe-C solid state phase transition, when I coupled the Mn element in the model, the following error appeared in the run log. I hope to get your help.
trying hard phases 2 3 level: 7 zp= 65758 error= 3
trying hard phases 2 3 level: 8 zp= 163858 error= 3
trying hard phases 2 3 level: 5 zp= 74298 error= 3
trying hard phases 2 3 level: 3 zp= 74298 error= 3
trying hard phases 2 3 level: 8 zp= 53752 error= 3
trying hard phases 2 3 level: 7 zp= 63800 error= 3
trying hard phases 2 3 level: 7 zp= 205031 error= 3
trying hard phases 2 3 level: 7 zp= 163762 error= 3
trying hard phases 2 3 level: 4 zp= 163762 error= 3
trying hard phases 2 3 level: 9 zp= 58974 error= 3
trying hard phases 2 3 level: 9 zp= 30391 error= 3
trying hard phases 2 3 level: 6 zp= 30391 error= 3
trying hard phases 2 3 level: 9 zp= 152717 error= 3
trying hard phases 2 3 level: 8 zp= 738 error= 3
trying hard phases 2 3 level: 8 zp= 738 error= 3
Re: solid-solid phase transformation
Dear kaiyang,
More information is required. Please attach input files or send them as PM if confidential.
More information is required. Please attach input files or send them as PM if confidential.
Re: solid-solid phase transformation
Dear Bernd,
Thank you very much for your reply. The previous problem has been solved.
Thank you very much for your reply. The previous problem has been solved.
Re: solid-solid phase transformation
Dear Bernd,
We are currently planning to simulate a non-diffusion phase change and want to prevent the elements from diffusing during the phase change. However, I found that there seems to be no diffusion-free option in the diffusion chapter of the MICRESS driver file. Is there any good solution to this problem?
We look forward to hearing from you.
We are currently planning to simulate a non-diffusion phase change and want to prevent the elements from diffusing during the phase change. However, I found that there seems to be no diffusion-free option in the diffusion chapter of the MICRESS driver file. Is there any good solution to this problem?
We look forward to hearing from you.
Re: solid-solid phase transformation
Dear kaiyang,
Whether a non-diffusion phase transformation occurs is not a problem of the phase properties (diffusivity), but of the relation between the interface velocity and the diffusion coefficients. MICRESS will always reproduce such "massive" transformations without segregation, if the driving force is high enough for overrunning any composition pile up, in a not diffusion but interface mobility controlled way. However, the physical question is whether the transformation front is so fast that even redistribution between the phases is impossible, o whether there is redistribution, but overrunning of the very slim composition peak. This makes a difference in the required driving force for the transformation.
As diffusion fields under these conditions are smaller than the interface thickness we typically use, and because MICRESS uses local (quasi-)equilibrium as standard, we cannot expect MICRESS to automatically switch between the two conditions and to get a quantitatively correct result, especially given the fact that the numerical interface thickness typically is not identical to a physically relevant one.
Therefore, in MICRESS, the user him/herself for such fast transformations has to decide between the two options, using "redistribution_control" and the options "nple" or "para" for each element. While the first assumes redistribution and takes the altered driving force due to the element pile-up into account, the second does not. Using "normal" or no redistribution control instead will lead to sort of "mixed" behavior which is purely numerical. There will then also be a risk of interface destruction due to strong gradients of composition.
Bernd
Whether a non-diffusion phase transformation occurs is not a problem of the phase properties (diffusivity), but of the relation between the interface velocity and the diffusion coefficients. MICRESS will always reproduce such "massive" transformations without segregation, if the driving force is high enough for overrunning any composition pile up, in a not diffusion but interface mobility controlled way. However, the physical question is whether the transformation front is so fast that even redistribution between the phases is impossible, o whether there is redistribution, but overrunning of the very slim composition peak. This makes a difference in the required driving force for the transformation.
As diffusion fields under these conditions are smaller than the interface thickness we typically use, and because MICRESS uses local (quasi-)equilibrium as standard, we cannot expect MICRESS to automatically switch between the two conditions and to get a quantitatively correct result, especially given the fact that the numerical interface thickness typically is not identical to a physically relevant one.
Therefore, in MICRESS, the user him/herself for such fast transformations has to decide between the two options, using "redistribution_control" and the options "nple" or "para" for each element. While the first assumes redistribution and takes the altered driving force due to the element pile-up into account, the second does not. Using "normal" or no redistribution control instead will lead to sort of "mixed" behavior which is purely numerical. There will then also be a risk of interface destruction due to strong gradients of composition.
Bernd