Inquiry on Coarsening Mesh Inside Grains During Phase-Field Simulation with EBSD Input in MOOSE

[PengWei97]

Dear MOOSE Team,

I hope this message finds you well.

I am currently working on a phase-field simulation that uses EBSD data as input to initialize a polycrystalline structure via the EBSDMeshGenerator and PolycrystalEBSD modules. The EBSD file I am using has a relatively dense mesh, and I would like to improve the computational efficiency by coarsening the mesh inside the grains during the simulation — without modifying the original EBSD file.

To address this, I enabled the mesh adaptivity system in MOOSE with the following settings:

[Executioner]
[./Adaptivity]
initial_adaptivity = 1
refine_fraction = 0.8
coarsen_fraction = 0.05
max_h_level = 1
[../]

However, in my current setup, I observed that while the grain boundary regions are successfully refined, the grain interiors remain at their original high resolution, and no coarsening occurs within them. This limits the performance gains I hoped to achieve.

My questions are as follows:

1. Is it possible to coarsen the mesh inside the grain interiors during simulation, while still refining near grain boundaries, without changing the EBSD file?
2. If so, how should I configure the mesh adaptivity to achieve this effect? Should I define custom error indicators (e.g., based on the gradient of phase-field variables like gr_i) to guide refinement/coarsening?
3. Are there example setups or recommendations for this kind of EBSD-based polycrystal model with adaptive mesh refinement/coarsening?

Thank you very much for your time and support. I look forward to any advice or guidance you can provide.

Best regards,
Wei

[GiudGiud]

Hello

The pre_refine parameter of the EBSDMeshGenerator is meant for this

It will create a coarser mesh that can then be further refined. See the doco page for that MG:
https://mooseframework.inl.gov/source/meshgenerators/EBSDMeshGenerator.html

If so, how should I configure the mesh adaptivity to achieve this effect?

mesh adaptivity cannot coarsen elements that have never been that coarse. It can only coarsen elements that were refined at some point.

Are there example setups or recommendations for this kind of EBSD-based polycrystal model with adaptive mesh refinement/coarsening?

see modules/combined/examples/phase_field-mechanics/EBSD_reconstruction_grain_growth_mech.i for an example using the pre-refinement (uniform_refine should be pre_refine in that input)

[PengWei97]

Thank you for your answer. According to your suggestions and optimizing the input inl data, I have solved the problem raised. Thank you very much.

wei