FEMDcoupler FEMDcoupler is a pre-processing toolkit for generating coupled Finite Element (FE) and Molecular Dynamics (MD) models. It creates simulation-ready systems by integrating an MD core with a surrounding FE domain, supporting features like pre-cracks and periodic boundaries. The workflow is orchestrated through a series of Python scripts that use Abaqus for FE mesh generation and perform data manipulation on LAMMPS data files. Features FE & MD Model Generation: Prepares 3D models with a central MD region integrated into a surrounding FE continuum based on a single LAMMPS data file. Supports both rectangular and cylindrical FE domains. Meshing: Generates a 3D hexahedral mesh. The mesh density is configurable and adapts to the geometry specified and given by the LAMMPS data file. Fracture Mechanics: Optionally introduces pre-cracks into the model. Notches: Create notches starting from the outer surface/edge with configurable width, depth, and orientation. Cutouts: Insert internal voids or channels (elliptical or cuboid) through the entire model. Periodic Boundary Support: Post-processes the Abaqus-generated mesh to ensure node-for-node periodicity, which allows for applying periodic Dirichlet boundary conditions (dPBCs). FE-MD Interface Handling: Anchor Atoms: Probabilistically places anchor atoms in the bridging region between the FE and MD domains to facilitate coupling. Interaction Management: Can automatically remove interactions at the boundary of the bridging region to avoid interactions between particles at non-periodic boundaries. Chain Truncation: Optional truncation of molecular chains at the FE-MD boundary to ensure clean coupling interfaces. Molecule Cleanup: Automated removal of molecules or fragments whose atom count is at or below a defined threshold, regardless of their origin (pre-existing or formed by other operations like cuts/notches/truncation). DPD Support: Capabilities to re-type atoms near non-periodic boundaries, facilitating Dissipative Particle Dynamics (DPD) setups. Configuration: A centralized Python configuration file (FEMDcoupler_params.py) allows for customization of the model, from geometry and mesh size to pre-crack parameters. This script generates a JSON configuration file used by the subsequent pipeline. Software Requirements Abaqus/CAE (2023 or older): Required for FE part creation and meshing. The scripts are designed to interface with the Abaqus Python environment. Note: The Abaqus Learning Edition (Student Version) is fully sufficient for using this toolkit as node limitations of the Learning Edition only apply to solving analysis jobs, not to generating the model files and meshes used here. Python 3.x: A standard Python 3 interpreter is needed to run the data processing and orchestration scripts. Installation Clone or download the repository to your local machine. Ensure that the abaqus command is available in your system's PATH. Alternatively, you can specify the full path to the Abaqus executable directly inside the FEMDcoupler_run.sh (Linux/macOS) or FEMDcoupler_run.bat (Windows) file. On Linux/macOS, make the shell script executable before the first run: chmod +x FEMDcoupler_run.sh Usage The primary workflow is managed by the main run scripts, which execute a sequence of Python scripts in the correct order. 1. Configuration All model parameters are defined within the ParamConfig class in the FEMDcoupler_params.py file. Before running the workflow, modify this file to define your desired model. Key configurable aspects include: FE domain geometry (rectangular or circular). Meshing and cell set parameters, as well as bridging region thickness. Cut and notch parameters for fracture studies. Anchor atom placement probability. Input and output file paths. Refer to the comments in FEMDcoupler_params.py for a detailed explanation of each parameter. 2. Execution On Windows: .\FEMDcoupler_run.bat On Linux/macOS: ./FEMDcoupler_run.sh Running these scripts will initiate the full pre-processing pipeline, culminating in the generation of the final coupled model files. Tips & Error Handling Running Scripts Individually: All Python scripts can be run independently. For a full process, simply follow the execution order stated under "Workflow Overview" below, which is also executed by the .bat or .sh files. Abaqus GUI Interaction: When the create_part_mesh.py script is executed, it will launch Abaqus/CAE. You can make arbitrary changes to the model within the GUI at this stage. Note for Linux Users: In some Linux environments, the script might not execute automatically once the GUI opens. If the model is not generated automatically, inside Abaqus go to File > Run Script... and select create_part_mesh.py manually. To ensure your changes are recognized by subsequently executed scripts, you must save the model by overwriting the already generated .inp file. To do so, expand Jobs in the Analysis section on the Model Worktree (click + left of it) > right click the only existing job listed there > Write Input > OK > Yes. If no changes are made, saving is not required, as the script handles file generation automatically. Abaqus may be closed without saving. Subsequent overwriting of .inp file and using periodic Dirichlet boundary conditions: If you overwrite the generated .inp file with Abaqus Job writing after the automated workflow has finished, it is recommended to manually rerun create_periodicity_mesh.py. This ensures that minor mesh irregularities introduced by the Abaqus mesher by saving/overwriting the .inp file are corrected. This is only relevant if you aim to use periodic Dirichlet boundary conditions in your following simulation. Abaqus Version Incompatibilities: create_part_mesh.py is written in Python 2.7, which is only compatible with Abaqus/CAE 2023 or older. For running it in Abaqus CAE 2024 or newer, please refer to this guide: https://tecnodigitalschool.com/upgrade-from-python2-to-python3-abaqus-2024/ Common Linux Issues: X2Go Rendering Issues: If running Abaqus/CAE via a remote desktop solution like X2Go, you may encounter rendering problems. To resolve this, add the -mesa flag after abaqus cae in the FEMDcoupler_run.sh for starting Abaqus. Headless Mode (recommended): To prevent the Abaqus GUI from starting altogether (e.g. to avoid graphics driver issues), you can modify the run script FEMDcoupler_run.sh as follows: After the abaqus cae command, replace script= with noGUI=. This is significantly faster and ensures the script executes automatically, whereas the GUI mode (script=) on Linux often requires triggering the script manually (see section Abaqus GUI Interaction). However, this removes the ability to interactively modify the FE model within the scope of the automated workflow. If the FE model/.inp file is modified later using Abaqus, please see section Subsequent overwriting of .inp file and using periodic Dirichlet boundary conditions. Workflow Overview The FEMDcoupler toolkit follows this sequence to build the final model: FEMDcoupler_params.py: Reads the user-defined parameters and generates a .json configuration file. create_anchors_cut_notch.py: Modifies the source LAMMPS data file based on the JSON configuration to introduce cuts, notches, and anchor atoms as well as to truncate atom bonds at the boundaries and to re-group boundary atoms for DPD deployment. create_part_mesh.py: Launches Abaqus/CAE to build the FE part, partition it, generate the mesh, assign sets, and write an initial .inp file. create_periodicity_mesh.py: Post-processes the .inp file, adjusting node coordinates to ensure perfect periodicity for dPBCs. create_merge_FEMD_data_inp.py: Merges the FE mesh data from the (corrected) .inp file with the modified MD atom data for inspection of the coupled configuration with tools like OVITO. License This software is distributed under the MIT License. See the LICENSE file for more details.