LSDYNA Material Modeling Project

Advanced framework for biaxial tensile test simulations and material model development using LS-DYNA, machine learning, and numerical analysis.

Features

• 🧪 Biaxial tensile test specimen modeling
• ⚙️ FEM analysis automation with LS-DYNA
• 🤖 Machine learning integration for material parameter calibration
• 📊 Experimental data processing pipeline
• 📈 Yield surface visualization and analysis
• 🔄 Automated batch processing workflows

Requirements

• Python 3.8+
• NumPy
• pandas
• Matplotlib
• scikit-learn
• Jupyter Notebook (for ML analysis)

Quick Start

pip install numpy pandas matplotlib scikit-learn

# Run tensile test analysis
python tensile_test.py

# Generate FEM models
python FEM_model_modify.py

# Start ML training
cd ML/nnc
python EX2NN.py

Project Structure

├── ML/ - Machine learning models and training pipelines
├── make_dataset/ - Data generation and preprocessing
├── yld2000/ - Yield function analysis and visualization
├── experiment_data/ - Raw experimental datasets
├── utils/
    └── script_helpers/ - Utility scripts and helper functions
└── analysis/ - Scripts for displacement and force analysis

Documentation

Dataset Generation Pipeline

Workflow Stages:

1. make_dataset/scripts/P1_cre_para_csv_3.py - Generates random material parameters using symbolic math
2. P2_cre_key_files.py - Creates LS-DYNA keyword files from parameters
3. P3_cre_outp_*.py - Standardizes FEM simulation outputs
4. P4_dyna_*.py - Automated LS-DYNA batch processing
5. P5_cre_inp_10.py - Prepares ML-ready input tensors
6. P6_*.py - Calculates yield surface errors
7. P7_edit_ex_modify_data3.py - Validates experimental data formats and converts to ML-compatible structures
8. make_dataset/scripts/P8_nnabla_run2.py - Neural network training execution with hyperparameter configuration

Key Features:

• 📂 Scripts organized in make_dataset/scripts and utils directories
• 🧩 Parameter space exploration with sympy
• ⚡ Parallel LS-DYNA execution
• 📈 Strain data normalization pipelines
• 🔄 ML dataset versioning support
• 🔍 Experimental data validation checks
• 🧠 Neural network configuration management

Material Calibration Tutorial | FEM Guide

Experimental Data Processing

• Tensile/bulge test analysis
• Strain distribution visualization
• Experimental/FEM data correlation

Machine Learning Integration

• Dataset split (train/val/test)
• Input feature engineering
• Output normalization layers
• Experimental data validation and preprocessing (P7)
• Neural network training workflows with hyperparameter configuration (P8)

Contributing

1. Fork the repository
2. Create feature branch (git checkout -b feature/amazing-feature)
3. Commit changes (git commit -m 'Add amazing feature')
4. Push to branch (git push origin feature/amazing-feature)
5. Open Pull Request

License

MIT

Material Modeling Project Setup

Virtual Environment

1. Activate virtual environment (Windows):

.\.venv\Scripts\activate

2. Install dependencies:

pip install -r requirements.txt

3. Deactivate with:

deactivate