Problems in simulating multiphase grain growth

[Buyu An]

Dear all

I'm learning to use MICRESS. When simulating grain growth during the sintering process of a system composed of two different solid phases (phase_1 = A and phase_2 = B) and a liquid phase (liquid), I encountered a problem: at time = 0, the simulation successfully generates two types of grains(grain type_1 = phase A, grain type_2 = phase B) . However, one of the solid phases, A, disappears completely in the next step, leaving only B in the simulated microstructure. Only the grains representing solid phase B participate in the subsequent simulations.

I would like to have both grains representing different solid phases coexist in the simulation, with one growing faster and the other slower.

According to the following output in the log：

tWidth_max( A: A) = 2.1375000E-02 s
tWidth_max( A: B) = 3.3398438E-02 s
tWidth_max( B: B) = 5.9375000E-02 s
Maximal value for tWidth = 2.1375000E-02 s for phase-field solver
Initial value for tWidth = 2.13750E-02 s for automatic time stepping (phase-field solver)
Critical grain radius:
of phase B in phaseA = 8.00000E-02 / dT_unt [micrometers]

Remaining license time: permanent

==================================================

Time t = 0.0000000 s
CPU-time: 3 s
Current phase-field solver time step = 2.14E-02 s
Temperature at the bottom = 1293.0 K
Temperature gradient = 0.00000 K/cm

**********************************************
* Beginning of simulation *
**********************************************

Phase 1 disappeared at 4.0092182E-03 s

I guess this may be related to the parameters between phase A and B in the phase interaction part.

What causes one of the solid phases, A, to disappear completely after the simulation begins?
How can I fix this?

Attached is my dri file. It is based on the T10_01_GrainGrowth_2D_dri file from the example.

Best wishes.
Buyu

[Bernd]

Dear Buyu,

Welcome to the MICRESS Forum.

Your input file you provide is not consistent, there is data missing. Nevertheless, the problem probably consists in the difference between the equilibrium temperature between phases A and B and the actual temperature. The overgrowth of phase A by phase B is so fast that it has already finished after 4.0092182E-03 s. To monitor what is happening, you need to write outputs after much shorter times, e.g.

linear_step 0.0001 0.005

The reason is the said temperature difference, which makes up the driving force for overgrowth, and the fact that you assume pure element phases (by not coupling to concentration). In reality, when you have an alloy with more than one element, diffusion in the solid would make the process much much slower, of course.

If you really want to have overgrowth of one phase by the other (like it happens now), you could adjust the time scale in the current setup by reducing the temperature difference with respect to the equilibrium temperature, or reduce the interface mobility of the A/B-interface. On the other hand, if you set the actual temperature to the equilibrium temperature, you will see only grain growth of both phases, which is happening in the expected time scale.

Currently, no liquid phase is present (at least according to how I reconstructed the missing parameters...). If you want to have it, you just can reduce the radii of the initial grains, so that they do not fully overlap.

Please don't hesitate to ask further questions if my answer was not conclusive or sufficient.

Best wishes
Bernd

[Buyu An]

Hi Bernd

Thank you so much for your patience and help. Thank you for pointing out the possible cause of the problem and providing suggestions.

I agree that the problem may be related to the equilibrium temperature between phases A and B. Solving this problem requires further clarifying the material system and the equilibrium temperature of its phases. I want to solve this problem after learning this software for a while.

My original goal was simply to observe the grain growth process of two grains with different growth rates. The purpose of defining different type grains as different phases is to set the interface migration parameters, because the simulation of grain growth is achieved through interface migration.

Therefore, when I set no_phase_interaction between phases A and B, set the number of the two types of grains to be the same, and have different interface migration parameters, I can also get the simulated growth results of different grains.

But I still have another question. When I run the example T00_03_Binary_E_HypoEutectic_2D_Lin in MICpad7.000, it gets stuck at

# Output
# ======
#
# Output times
# ------------
# Finish input of output times (in seconds) with 'end_of_simulation'
# 'regularly-spaced' outputs can be set with 'linear_step'
# or 'logarithmic_step' and then specifying the increment
# and end value
# ('automatic_outputs' optionally followed by the number
# of outputs can be used in conjuction with 'linear_from_file')
# 'first' : additional output for first time-step
# 'end_at_temperature' : additional output and end of simulation
# at given temperature
3.
linear_step 1. 100.
end_at_fraction 0 0 0.

The output log is shown as follows

Seed type 2 will be described by the nucleation model seed_undercooling
Maximum number of new nuclei of seed type 2: 10
Grain radius [micrometers]: 0.00000
Choice of growth mode: stabilisation
min. undercooling [K] (>0): 2.0000
Determination of grain orientations is 'random'
Shield time [s]: 2.0000
Shield phase or group number: 2
Shield distance [micrometers] [ nucleation distance [micrometers] ]: 10.000
Distance between simultaneous Nucleations [micrometers]: 50.000
min. nucleation temperature for seed type 2: 0.0000 K
max. nucleation temperature for seed type 2: 2000.0000 K
Time between checks for nucleation: 0.10000 [s]
No random noise will be applied.

Other nucleation parameters
---------------------------
Max. number of simultaneous nucleations: 4611686018427387913
Metastable small seeds will not be killed


Warning! Trailing characters in string: end_at_fraction 0 0 0.




Input error in routine leseEingabe
Input: end_at_fraction 0 0 0.

This should be a minor issue, but what causes it and how can I fix it?

Looking forward to your reply.
Best wishes
Buyu

[Bernd]

Dear Buyu,

You should not use "no_phase_interaction" between phases A and B,

Therefore, when I set no_phase_interaction between phases A and B, set the number of the two types of grains to be the same, and have different interface migration parameters, I can also get the simulated growth results of different grains.

because then, the interfaces between the different types of grains cannot move! You should rather set the temperature to the equilibrium temperature for A/B. So you avoid "chemical" interactions while allowing for grain growth.

The problem with example T00_03_Binary_E_HypoEutectic_2D_Lin comes from a typo in the input file which is not present in the file you attached. I don't know how you got it there, but it is definitively wrong:

end_at_fraction 0 0 0.

The "end_at_fraction" keyword only needs 2 parameters (the phase number and the critical phase fraction). Perhaps, the problem has to do with using MICpad 7.000, because the example T00_03_Binary_E_HypoEutectic_2D_Lin is newer (MICRESS 7.1 and later). Which MICRESS version do you use currently? You should always use the same versions of MICRESS, MICpad and the corresponding Training Examples.

Bernd

[Buyu An]

Dear Bernd，

Thank you very much for your advice!

Yes, the interface between the different types of grains cannot move. I will follow your suggestion, set the temperature to the equilibrium temperature for A/B, and run the simulation again.

I'm using MICpad version 7.00. But the example T00_03_Binary_E_HypoEutectic_2D_Lin is from the 7.3 examples, downloaded from

https://www.micress.de/download.html#examples

The versions are indeed different. I will download examples that match version 7.00.

Thank you again for your patience and help！
It has greatly encouraged me to learn this software. Your help gave me the confidence and enthusiasm to solve all the problems I will encounter.

Thank you！
Bernd

[ralph]

Be aware that MICpad does not run the simulation itself, i.e. you can run the examples from MICRESS 7.3 in MICpad 7.0 if you configure a MICRESS version 7.3 in the run configuration. MICpad just starts MICRESS and displays the output or tabular result files.

MICpad can show results calculated on Linux or Windows. These result files have different file extensions. When you like to rerun an examples, change the output location before or delete existing results. Otherwise you mix Linux and Windows results and MICpad might show you the wrong files.

Best,
Ralph