Damask version including KWN model for precipitation kinetics

This version of Damask includes a constitutive behaviour law for dynamic precipitation. The program is available here.

The constitutive law called kwnpowerlaw allows to consider the evolution of a precipitate distribution under deformation. The evolution of the precipitate distribution is calculated with a multi-class KWN precipitation kinetics model. The strengthening effect of precipitates is taken into account in the calculation of the critical resolved shear stress for dislocation glide. The accelerating effect of deformation on precipitation kinetics is considered using a phenomenological model for excess vacancy production.

The documentation of the model, which describes the equations used and the corresponding entries in the input files, can be downloaded here

Installation

The following installation instructions are specific to users of the University of Manchester’s Computational Shared Facility (CSF).

First connect to the CSF
if it does not already exist, create a software folder in your home by running the following command:

mkdir software

Download the files load_DAMASK.sh and load_DAMASK_processing.sh from the env-script folder (here) and copy them in your home folder
In your home directory, run the following commands:

source load_DAMASK_processing.sh

Download the folder called damask-kwn from this link and copy it in your software directory (or clone it directly from Github)
Go to the software/damask-kwn/ directory:

cd software/damask-kwn/

Go to the src directory and remove the hidden files that might have appeared when copying the damask-kwn folder in the software directory (this step is not necessary if you used github to clone the folder):

cd src
rm ._*
cd ..

Create a build directory in software/damask-kwn/ and navigate into it:

mkdir build

cd build

Run the following command:

cmake ../ -DDAMASK_SOLVER=GRID -DCMAKE_INSTALL_PREFIX=../

Run the following command

make all install

Running simulations

First of all, open the CSF and run the following command:

source /mnt/eps01-rds/jf01-home01/shared/load_DAMASK_processing.sh

Running a job

Place the following jobscript.sh file into a directory containing DAMASK input files (geom.vtr, load.yaml, and material.yaml) and submit it with

qsub jobscript.sh

where jobscript.sh looks like this

#!/bin/bash --login
#$ -cwd                   # Submit in the current working directory
#$ -pe smp.pe 4           # Use a parallel environment with four cores

source ~/load_DAMASK.sh

mpirun -n $NSLOTS DAMASK_grid -l load.yaml -g geom.vtr

Running a job on the CSF will create two files in the working directory it is run within: A jobscript.sh.o0000000 file, which contains generic job output, and jobscript.sh.e0000000 which contains detail on errors that occured during the run.

A set of examples files (geom.vtr, load.yaml, material.yaml and jobscript.sh files) is available in the example_of_use directory here

Input files examples

The material.yaml file contains all the material properties. It also contains all the parameters of the KWN dynamic precipitation model (please note that the temperature is defined in the load.yaml file). The documentation detailing the meaning of the inputs is available here.

Example material.yaml file (for a simulation with 4 grains)

homogenization:
  SX:
    N_constituents: 1
    mechanical:
      type: pass
    thermal:
      type: pass
      
material:
- constituents:
  - O:
    - 0.19813654684736873
    - -0.4003702989793671
    - 0.38708148638970563
    - -0.806606133991621
    phase: Aluminum
    v: 1.0
  homogenization: SX
- constituents:
  - O:
    - 0.4131826988800702
    - -0.6684243946439803
    - 0.5135613787035137
    - 0.34459192720543513
    phase: Aluminum
    v: 1.0
  homogenization: SX
- constituents:
  - O:
    - 0.21956653913118745
    - -0.3396582332002957
    - 0.5011663510132476
    - -0.765019678260156
    phase: Aluminum
    v: 1.0
  homogenization: SX
- constituents:
  - O:
    - 0.2487765086598615
    - -0.8408812909074685
    - 0.44711671323124785
    - 0.17639599794237146
    phase: Aluminum
    v: 1.0
    
    
phase:
  Aluminum:
    lattice: cF
    mechanical:
      elastic:
        C_11: 100000000000.0
        C_12: 60410000000.0
        C_44: 28340000000.0
        type: Hooke
      output:
      - F
      - P
      - F_e
      - F_p
      - L_p
      - O
      plastic:
        N_sl:
        - 12
        a_sl: 2.25
        atol_precipitate_density: 1e-12
        atol_solute: 1e-12
        atol_xi: 1.0
        atomic_volume: 1.66e-29
        burgers_vector: 2.9e-10
        c0_matrix:
        - 0.02889
        - 0.02406
        ceq_matrix:
        - 0.012082196019601959
        - 0.0005296453049830318
        ceq_precipitate:
        - 0.27
        - 0.38
        dislocation_arrangement: 1
        dot_gamma_0_sl: 0.001
        gamma_coherent: 0.265
        h_0_sl_sl: 589020809.2851744
        h_sl_sl:
        - 1
        - 1
        - 1.4
        - 1.4
        - 1.4
        - 1.4
        - 1.4
        initial_dislocation_density: 100000000000000.0
        initial_mean_radius: 9.0e-10
        initial_volume_fraction: 0.007
        jog_formation_energy: 4.806529901999999e-20
        kwn_nsteps: 60
        kwn_step0: 5
        kwn_stepsize: 0.5
        lattice_parameter: 4.0695e-10
        misfit_energy: 0
        molar_volume: 1e-05
        n_sl: 50.0
        output:
        - gamma_sl
        - phi_total
        - phi
        - solute_c
        - r_avg
        - xi_sl
        - vacancy_c
        - f
        precipitate_strength_constant: 0.03535
        transition_radius: 3.3e-9
        shear_modulus: 28340000000.0
        solute_diffusion0: 1.49e-05
        solute_migration_energy: 2.0923e-19
        solute_strength: 683000000.0
        standard_deviation: 2.01322e-10
        stoechiometry:
        - 5
        - 7
        - 6
        type: kwnpowerlaw
        vacancy_diffusion0: 1e-05
        vacancy_energy: 8.331e-20
        vacancy_generation: 0.035
        vacancy_migration_energy: 1.49e-19
        vacancy_sink_spacing: 5e-05
        xi_0_sl:
        - 7800000.0
        xi_inf_sl:
        - 210330547.0653053
    rho: 1.0
    thermal:
      C_p: 1.0

The load.yaml file contains the deformation conditions as well as the temperature.

Example load.yaml file

initial_conditions:
  thermal:
    T: 423.0
loadstep:
- boundary_conditions:
    mechanical:
      P:
      - x
      - x
      - x
      - x
      - 0
      - x
      - x
      - x
      - 0
      dot_F:
      - 0.0001
      - 0
      - 0
      - 0
      - x
      - 0
      - 0
      - 0
      - x
  discretization:
    N: 10000 #discretisation for the calculation (N time steps)
    t: 2000 #total deformation time in [s]
  f_out: 100 #frequency of the outputs : one file every f_out calculations 

  
solver:
  mechanical: spectral_basic
  thermal: spectral

Post-processing

Some post-processing tools are available in the env_scripts directory from this repository. To use them, copy the post_processing.py and notebook_post_processing.ipynb in the folder containing the result file (which has .hdf5 extension).

To create textfiles containing the results of the calculation, as well as .vtr files allowing to display the result in 3D (e.g. with Paraview), go with the CSF to the folder containing the result file (geom_load.hdf5 in the example below) and run the following command:

python post_processing.py geom_load.hdf5

This will create a textfile per variable and per time increment containing the value of the given variables for all voxels of the simulation. For example, if the simulation is run with a 16x16x16 (=4096) box, the file vf_100 contains 4096 lines for the calculated volume fractions of the 4096 elements of the simulation at increment time 100. This also creates .vtr files that can be visualised in Paraview.

Additionally, the jupyter notebook notebook_post_processing.ipynb can be copied in the folder containing the result file and run to display some plots.