run_sim

czmtestkit.py_modules.run_sim(name, doe_data, fixed_data, abaqus_simFunc=None, abaqus_postProc=None, postProc=None)

Sequentially run Abaqus/CAE simulations and/or post processing functions for a design of experiments.

Parameters

name (str): ID for colleciton of tests in the design of experiments.

doe_data (dict): dictionary with variables for tests in the design of experiments.

‘nPoints’

List of indicies corresponding to the tests to be run. (The defintion of this attribute was changed, see version history for details.)

‘Variable_Key_1’

List of values for the variable named Variable_Key_1.

‘Variable_Key_2’

List of values for the variable named Variable_Key_2.

fixed_data (dict): dictionary with constants for tests in the design of experiments.

‘Constant_Key_1’

Value of the variable named Constant_Key_1.

‘Constant_Key_2’

Value of the variable named Constant_Key_2.

abaqus_simFunc (str): Name of abaqus-python function from czmtestkit.abaqus_modules. See Example for the differnce between abaqus_simFunc and abaqus_postProc parameters and czmtestkit.abaqus_modules for available functions and instructions to create your own abaqus-python function that is compatible with the czmtestkit.

abaqus_postProc (str): Name of abaqus-python function from czmtestkit.abaqus_modules. See Example for the differnce between abaqus_simFunc and abaqus_postProc parameters and czmtestkit.abaqus_modules for available functions and instructions to create your own abaqus-python function that is compatible with the czmtestkit.

postProc (function object): Executable python post processing function.

Example

Assume that abqPy_Func1 is a function in the czmtestkit based on abaqus-python scripting language to run Abaqus/CAE simulation. Also assume that abqPy_Func1 takes a dictionary as input parameter with param_1, param_2 and param_3 as keys within the dictionary.

In_dict_Func1 = {
    'param_1': value_param1,
    'param_2': value_param2,
    'param_3': value_param3,
}

If param_1 requires a string input then value_param1 has to be a string and so forth. Further, if the design of experiments (doe) is such that the param_2 is to assume different values:

doe_param2 = [value0_param2, value1_param2, value2_param2, value3_param2]

and the aim is to run the simulations for the first and the third points (value0_param2, value2_param2) in this doe, then define the input dictionaries as follows:

dict_fix = {
    'param_1': value_param1,
    'param_3': value_param3,
}

dict_var = {
    'param_2' : doe_param2,
    'nPoints' : [0,2]
}

Use the run_sim() function to sequentially run the tests:

run_sim('ExampleDOE', dict_var, dict_fix, abaqus_simFunc="czmtestkit.abaqus_modules.abqPy_Func1")

This creates the main directory ExampleDOE in the current work directory and sub directories for each test. Further, with sub directories as working directories, each test in the nPoints is executed resulting in abaqus files.

$ <current working directory>
└── ExampleDOE
    ├── point_00
    │   ├── <abaqus file 1 test 1>
    │   ├── <abaqus file 2 test 1>
    │   ├── <abaqus file 3 test 1>
    │   ├── <abaqus file 4 test 1>
    │   ├── <abaqus file 5 test 1>
    │   └── point_00.json
    └── point_02
        ├── <abaqus file 1 test 2>
        ├── <abaqus file 2 test 2>
        ├── <abaqus file 3 test 2>
        ├── <abaqus file 4 test 2>
        ├── <abaqus file 5 test 2>
        └── point_02.json

Here, the first test is exectued with the following dictionary saved to point_00.json and passed to czmtestkit.abaqus_modules.abqPy_Func1() as input.

{
    'param_1': value_param1,
    'param_2': value0_param2,
    'param_3': value_param3,
}

and the input dictionary for the second test with the third point in the doe saved to point_02.json as

{
    'param_1': value_param1,
    'param_2': value2_param2,
    'param_3': value_param3,
}

Remaining tests in the doe can be executed by adding corresponding indices to the ‘nPoints’ list in the doe_data. Further, if a postprocessing function abqPy_Func2 has to be executed with parameters param_1, param_2 and a new parameter param_4. run_sim() can be used again.

dict_fix2 = {
    'param_1': value_param1,
    'param_4': value_param4,
}

dict_var2 = {
    'param_2' : doe_param2,
    'nPoints' : [0,1,2,3] # assuming `abqPy_Func1` has already been executed for all the points. The directories should have already been created.
}

run_sim('ExampleDOE', dict_var2, dict_fix2, abaqus_postProc="czmtestkit.abaqus_modules.abqPy_Func2")

Similarly, python postprocessor (py_func) can be executed sequentially for the doe using run_sim(...,postProc=py_func). Executing a postProc function results in the creation of a Database.json in the test directory with input dictionaries from all the tests.

$ <current working directory>
└── ExampleDOE
    ├── point_00
    ├── point_01
    ├── point_02
    ├── point_03
    └── Database.json

Database.json

{'param_1': value_param1, 'param_2': value0_param2, 'param_3': value_param3, 'param_4': value_param4 }
{'param_1': value_param1, 'param_2': value1_param2, 'param_3': value_param3, 'param_4': value_param4 }
{'param_1': value_param1, 'param_2': value2_param2, 'param_3': value_param3, 'param_4': value_param4 }
{'param_1': value_param1, 'param_2': value3_param2, 'param_3': value_param3, 'param_4': value_param4 }

If the py_func results in an output dictionary {'output1': <value>, 'output2': <value>}, the contents of the output dictionary are appended to the input dictionaries before writting to the Database.json.

Database.json

{'param_1': value_param1, 'param_2': value0_param2, 'param_3': value_param3, 'param_4': value_param4, 'output1': value0_out1, 'output2': value0_out2}
{'param_1': value_param1, 'param_2': value1_param2, 'param_3': value_param3, 'param_4': value_param4, 'output1': value1_out1, 'output2': value1_out2}
{'param_1': value_param1, 'param_2': value2_param2, 'param_3': value_param3, 'param_4': value_param4, 'output1': value2_out1, 'output2': value2_out2}
{'param_1': value_param1, 'param_2': value3_param2, 'param_3': value_param3, 'param_4': value_param4, 'output1': value3_out1, 'output2': value3_out2}

Note

All three functions can be executed at once.

FixDict = {
    'param_1': value_param1,
    'param_3': value_param3,
    'param_4': value_param4,
}

VarDict = {
    'param_2' : doe_param2,
    'nPoints' : [0,1,2,3]
}

run_sim('ExampleDOE', VarDict, FixDict, abaqus_simFunc="czmtestkit.abaqus_modules.abqPy_Func1", abaqus_postProc="czmtestkit.abaqus_modules.abqPy_Func2", postProc=py_func)

Metadata

Environment

Python

Version

v1.1.0	Updated type and functionality of doe_data[‘npoints’]. v1.0.0: (Int) Number of points in the design on experiments.
v1.0.0	base version

Date

2021-12-11

Authors

Nanditha Mudunuru

Contribution: v1.0.0, v1.1.0

Email: nanditha.mudunuru@gmail.com