Treatment of interface area in MICRESS

[ilovemicress]

Hi Bernd,

I have a question about ferrite and austenite phase fractions.
Would you like to explain how MICRESS calculates the fractions?
In “tabF.txt”, the summation of the fractions of the ferrite and austenite phases is always 1 regardless of existence of interface areas. Please refer an attached file.

Regards,
Taka

[Bernd]

Hi Taka,

of course, the sum of the ferrite and austenite fractions must be 1 if no other phase is present! At the interface region you have a phase mixture with the local phase fractions being identical to the phase-field parameter. The .TabF output is obtained by integrating the phase-field paramter of all grains of the corresponding phase.
In the .phas output, only the regions with above 70% fraction of the locally present grain are marked with the corresponding phase color, other regions with fraction 30-70% are blue. This is just to visualize grain boundaries. Thus, the total phase fractions cannot be obtained from the .phas output! You can just see the identity of the majority phase!
By the way, the exended blue regions in your case indicate a "spreading" of the interface region, i.e. a destroyed interface profile. This can be due to a too low resolution and/or too high mobility.
Bernd

[ilovemicress]

Hi Bernd,

Thank you for your explanation.
I have understood about the treatment of interface area in MICRESS.

Are there techniques which change a resolution and/or mobility with maintaining the fractions and shapes of each phase?

Regards,
Taka

PS: Can I change the threshold, that is “30-70%”, to arbitrary values?

[Bernd]

Hi Taka,

actually there is no possibility to change this default threshold in the .phas output. But you can use the optional parameter "no_interfaces" to remove the blue region and to always display the mayority phase. Another possibility would be to create an individual output with DP_MICRRESS using the .phas and .frac(x) outputs (see here).

Changing resolution is very easy: You just alter the cell dimension and the grid size (number of cells) in the geometry input. If there are no resolution-related artifacts, this should not change phase fractions.

Changing mobility should (in the ideal case) also not change the phase fractions if the considered phase transformations are diffusion-limited and resolution is high enough. Calibration of mobility in case of unsufficient resolution is explained here.

Bernd

[ilovemicress]

Hi Bernd,

Thank you for your advice.
I referred the threads and they were helpful.

Now, I am reading the new manual of MICRESS 6.0.
I will ask you other questions later.

Regards,
Taka