Error 402 followed by Error 100 in DAMASK simulation

[LeoDanninger]

Dear colleagues,

I’m currently facing an issue while running simulations using DAMASK. Briefly, I’m applying a tensile load to a sample at a fixed strain rate.

In my initial tests with a relatively small grid (16×16×16), the simulation ran without any problems. However, when I increased the grid size to 128×128×128, I encountered Error 402. After searching this forum, I found that this error might be related to large deformations, and that enabling regridding could be a potential solution (#50).

Although my deformation is not very large (engineering strain = 0.12), I tried enabling regridding. Unfortunately, when doing so, I now get Error 100 instead.

Has anyone experienced a similar issue or have suggestions on how to solve this?

Below are the commands I used to launch the simulations:

mpiexec -n 16 DAMASK_grid -j TiTeste2 -l tensionX_1e-3_f_out=625.yaml -g poly_Ti_100_128x128x128.vti -m material.yaml 2>&1 | tee Teste2.log
mpiexec -n 16 DAMASK_grid -j TiTeste2 -l tensionX_1e-3_f_out=625.yaml -g poly_Ti_100_128x128x128.vti -m material.yaml -r 625 2>&1 | tee Teste2.log

The input files for the simulation, along with the log files containing the error messages, are attached below for your reference:
Input_1e-3.zip

Note 1: I used a high f_out value because I'm only interested in a few output points, and the file size was getting too large with lower values.
Note 2: The 2>&1 | tee Teste2.log part is only to store the log in a file — I still get the same errors even without this logging command.

Any help or insight would be greatly appreciated!

Best regards,

Leo R D

[ruanqc]

hi, Leo,
in your error 100

 │ could not open file:                                                │
 │ TiTeste2_restart.hdf5 

which indicates your missing your restart output file "TiTeste2_restart.hdf5", for restart you should defined restart output freequncy through " f_restart: N_r in load yaml file, eg.,

 solver: {mechanical: spectral_basic}
loadstep:
- boundary_conditions:
    mechanical:
      dot_F:
      - [0.001, 0, 0]
      - [0, x, 0]
      - [0, 0, x]
      P:
      - [x, x, x]
      - [x, 0, x]
      - [x, x, 0]
  discretization: {t: 125, N: 1250, r: 1.0}
  f_out: 625
  f_restart: 625  # evry N_restart increment to write to the restart file

when run the simulation command:
mpiexec -n 16 DAMASK_grid -j TiTeste2 -l tensionX_1e-3_f_out=625.yaml -g poly_Ti_100_128x128x128.vti -m material.yaml

which will also output another hdf5 file called "TiTeste2_restart.hdf5" along side with "TiTeste2.hdf5" for storing restart information, then rerun simulation, (see more in #71)
command:
mpiexec -n 16 DAMASK_grid -j TiTeste2 -l tensionX_1e-3_f_out=625.yaml -g poly_Ti_100_128x128x128.vti -m material.yaml -r 625 2>&1 | tee Teste2.log

one IMPORTANT thing for restart is that make sure the latest output increment in TiTeste2.hdf5 not exceed N_r ( for instance 625 in above case), otherwise will called error shown there already has and can not restart properly. (see more in #35)

Hope it helps and best regards, ruanqc

[LeoDanninger]

Dear ruanqc,

I apologize for the delay in getting back to you.

I followed your suggestions and was able to successfully restart the simulation without any issues. I truly appreciate your help.

I’m still trying to understand the root cause of the initial error (error 402), especially considering that the model runs without crashing for grid sizes up to 64x64x64, even at relatively low strain levels. However, regarding error 100, your explanation was extremely helpful, and I now have a much better understanding of how to properly restart a simulation.

Once again, thank you very much.

Best regards,

Leo R D

[ruanqc]

Dear ruanqc,

I apologize for the delay in getting back to you.

I followed your suggestions and was able to successfully restart the simulation without any issues. I truly appreciate your help.

I’m still trying to understand the root cause of the initial error (error 402), especially considering that the model runs without crashing for grid sizes up to 64x64x64, even at relatively low strain levels. However, regarding error 100, your explanation was extremely helpful, and I now have a much better understanding of how to properly restart a simulation.

Once again, thank you very much.

Best regards,

Leo R D

@LeoDanninger : Glad you have solve the restart problem. I am not familiar with your error 402, but i think its related with the orientation matrix, see src/rotation.f90 (Line 234 )

 if (dNeq(math_det33(om),1.0_pREAL,tol=1.0e-5_pREAL)) &
call IO_error(402,ext_msg='fromMatrix')

maybe some cells are distorted too much so it can not properly convert to rotation matrix.