New grain formation using Grain tracker

[ritamcj]

Dear Moose users,
I am trying to simulate nucleation in a polycrystal using Grain tracker. I came across the option "reserved order parameters" in Grain tracker but I am unable to figure out how to invoke new grain formation using this. I tried various strategies as follows:

1. I used discretenucleation to increase new grain order parameters (say gr3,gr4,gr5) to values greater than the threshold required to recognise a new feature. I kept op_num = 6, initial grains = 3 and used polycrystalvariables and polycrystalkernel to evolve grain op's with prefix "gr".

[./nucleation_energy]
  type = DiscreteNucleation
  f_name = Fn
  op_names  = 'gr0 gr1 gr2 gr3 gr4 gr5'
  op_values = '0.1 0.1 0.1 1 1 1'
  penalty = 20
  penalty_mode = MATCH
  map = map
  block = 'face1_QUAD4  face2_QUAD4  face3_QUAD4'
  outputs = exodus
[../]

2. Increase the number of order parameters "op_num" in my input file by writing a simple function. Something like:

[GlobalParams]
  op_num = op_evolution_function
  deformed_grain_num = 3
  var_name_base = gr
  grain_tracker = grain_tracker
[../]

[Functions]
  [./op_evolution_function]
    type = ParsedFunction
    value = deformed_grain_num + 1
  [../]
[]

This did not work as the syntax to read a function value is likely incorrect. I have come across functions in Grain tracker such as newGrainCreated, getNewGrainIDs but am unable to figure out how to invoke nucleation through new grain order parameters. I would really appreciate guidance on possible strategies to tackle this problem.

Warm Regards,
Ritam

[laagesen]

Hi Ritam,

If you haven't already, please have a look at the documentation page

https://mooseframework.inl.gov/modules/phase_field/Nucleation/DiscreteNucleation.html

and especially the section Direct Order Parameter Modification. You will need to add a DiscreteNucleationForce and Reaction kernel that act on the reserved order parameter. You shouldn't need to do either of the other things you tried, but you will need a map and inserter to inform the DiscreteNucleationForce kernel.

[dschwen]

Also make sure you run nucleation only on the reserved order parameters. Those are not assigned any grains by the grain tracker and can be use to stage new nuclei.

[ritamcj]

Hi Larry and Daniel,
Thank you so much for your suggestions. I am able to see nucleating grains individually through their order parameters (say gr9). I am now trying to observe the evolving grain structure, but the new nuclei aren't observed in "unique_grains" in GrainTracker even though the new grain order parameters exceed the threshold. However, the halos do update. Any pointers regarding this? I have applied a grain growth model through "polycrystalstoredenergy" and am able to observe grain growth in pre-existing grains.
Ritam

new_grains.mp4

unique_grains.1.mp4

halos.mp4

[laagesen]

Ritam, you will need to use a significantly higher mesh resolution than shown here. The GrainTracker is probably having a hard time distinguishing new grains from existing grains. The nucleation rate also seems to be very high so it may be having a hard time distinguishing between individual nucleation events. Please make sure you have at least 3 mesh elements through the interfaces between order parameters prior to trying anything with nucleation. Once you observe grain growth under these conditions, you can then add in nucleation.

[ritamcj]

Hi Larry,
I am happy to inform you that I am able to visualize the evolving microstructure with nucleating grains, which grow further with stored deformation energy based on dislocation density evolution. I used a very fine mesh and a slower nucleation rate and it worked. Many thanks for your suggestions.
Warm regards,
Ritam

new_grains.mp4

unique_grains.mp4

[laagesen]

Great! Glad that worked.

[Sabalei]

Hi Ritam,
I'm trying to simulate static recrystallization. I also applied a grain growth model through "polycrystalstoredenergy".
I have inserted the nucleus through Direct order parameter modification. However, new nuclei shrink after formation rather than grow further. I tried to increase the hold time and Dislocation density, but it didn't solve the problem. Any pointers regarding this?

[ritamcj]

Hi Sabalei,
You can increase the order modification parameter (v1) to increase the driving force for growth of nuclei. Please see https://mooseframework.inl.gov/source/kernels/DiscreteNucleationForce.html
Now, you should be able to see growth followed by shrinkage and annihilation of nuclei.
Warm regards,
Ritam

[laagesen]

Thanks @ritamcj!

[mdh23rct]

@ritamcj @Sabalei @GiudGiud Dear all,
Have you got any grain coalescence problem in your new grain formation simulation? I am finding two grains with the same grain order parameter are coalescing with each other (Please see the attached image). However, I would like to prevent the coalescence of grains and keep them sepearate with a grain boundary. Would you be kind to let me know what sort of parameters I should use in the GrainTracker userobject to make that happen? Your advice will be highly appreciated. Thanks.

[Sabalei]

Are these new grains formed by nucleation? Whether remap is enabled? Coalescence can be prevented by remap two grains that are very close together into different order parameters.

[mdh23rct]

@Sabalei @GiudGiud
Yes, new grains are forming by discrete nucleation. As you suggested, I have enabled the remapping. However, I am getting a
new problem. Duplicate grains are forming at the same location when remap_grains = true. Please see the attached image. Any solution for that? Thanks.