Gen-SHM: Physics-Informed Generative Surrogate for Drone Wing Structural Integrity

This repository implements a complete physics-informed generative surrogate model for drone wing damage detection using Physics-Informed Neural Networks (PINNs), with professional client-ready deployment.

🚀 Overview

The system generates synthetic vibration data for arbitrary damage scenarios in drone wings by embedding the Euler-Bernoulli beam equation into a parametric neural network. This addresses the critical lack of failure data for training structural health monitoring systems while providing mission-critical safety demonstrations.

Professional Features:

• ✅ Client-ready deployment with automated setup
• ✅ Cross-platform compatibility (Linux/macOS/Windows)
• ✅ Comprehensive diagnostic tools
• ✅ Multiple demonstration modes
• ✅ Robust error handling and graceful degradation

🌟 Key Features

Core Technology

• Physics-Informed Generation: Embeds Euler-Bernoulli beam theory with spatially varying stiffness
• Damage Parameterization: Models crack location and severity through stiffness reduction
• Real-time Capabilities: Lightweight surrogate suitable for edge deployment
• Zero-shot Detection: Enables damage identification without destructive testing
• Comprehensive Validation: Physics compliance checks and SHM performance metrics

Professional Deployment

• Automated Client Setup: One-command installation with setup_client.sh
• Multiple Demo Modes: Text-based, GUI, and diagnostic demonstrations
• Cross-Platform Support: Works on Linux, macOS, and Windows
• Robust Error Handling: Graceful degradation when components fail
• Comprehensive Diagnostics: Built-in troubleshooting tools

🛠️ Installation

Quick Start (Recommended)

cd gen-shm
./setup_client.sh

Manual Installation

cd gen-shm
python3 -m venv gen-shm-client-env
source gen-shm-client-env/bin/activate  # Linux/macOS
# gen-shm-client-env\Scripts\activate     # Windows
pip install -r requirements.txt

▶️ Quick Start

Interactive Mission Demo (Most Impressive)

python revolutionary_demo.py

Shows mission-critical scenario comparing protected vs unprotected drones

Quick Showcase Demo

python showcase_demo.py

Text-based demonstration of core functionality

Diagnostic Tool

python client_diagnostic.py

Identifies and fixes common deployment issues

Programmatic Usage

# Generate synthetic vibration data for damage detection
from src.models.surrogate_model import DroneWingSurrogate

# Initialize model
surrogate = DroneWingSurrogate()

# Generate 100 samples with 20% crack at wing root
samples = surrogate.generate_samples(
    damage_level=0.2,
    damage_location=0.0,  # Root location
    num_samples=100
)

📁 Complete Project Structure

gen-shm/
├── src/                    # Core source code
│   ├── models/            # Neural network architectures and physics
│   ├── data/              # Data generation and preprocessing
│   ├── training/          # Training loops and optimization
│   ├── evaluation/        # Metrics and validation
│   └── utils/             # Utility functions
├── experiments/           # Main experiment scripts
├── notebooks/             # Jupyter demonstrations
├── configs/               # Configuration files
├── tests/                 # Unit and integration tests
├── demos/                 # Interactive demonstrations
│   ├── showcase_demo.py          # Text-based showcase
│   ├── revolutionary_demo.py     # Mission-critical GUI demo
│   └── direct_working_demo.py    # Minimal output demo
├── deployment/            # Client deployment tools
│   ├── setup_client.sh           # Automated setup script
│   ├── launch_safe.py            # Robust launcher
│   ├── client_diagnostic.py      # Diagnostic tool
│   └── README_CLIENT.md          # Client documentation
└── documentation/         # Additional guides

Key Components

Physics Foundation

• beam_physics.py: Implements Euler-Bernoulli beam equation with damage parameterization
• Physics residual computation using automatic differentiation
• Boundary condition enforcement

Generative Architecture

• pinn_generator.py: Parametric PINN that learns solution operator across damage parameters
• Input: space-time coordinates (x,t) + damage parameter (d)
• Output: vibration response u(x,t)

Training Framework

• Hybrid loss combining data fidelity and physics compliance
• Adaptive loss weighting strategies
• Multi-scale training progression

🧪 Running Experiments

Main Training

python experiments/train_model.py --config configs/default.yaml

Sample Generation

python experiments/generate_samples.py --damage_level 0.15 --num_samples 50

Performance Evaluation

python experiments/evaluate_shm.py --model_path checkpoints/best_model.pt

Client Diagnostics

python client_diagnostic.py

Safe Launcher

python launch_safe.py

Provides menu-driven access to all demonstrations

📚 Documentation

• Client Deployment Guide - Complete setup and troubleshooting
• Getting Started - Developer setup guide
• Implementation Details - Technical architecture

🤝 Contributing

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

📖 Citation

If you use this code in your research, please cite:

@article{gen-shm2026,
  title={Gen-SHM: Physics-Informed Generative Surrogate for Drone Wing Structural Integrity},
  author={Saurav},
  year={2026}
}

📄 License

MIT License

🆘 Support

For issues, questions, or contributions:

• Check the Client Deployment Guide first
• Run python client_diagnostic.py for automatic troubleshooting
• Review the Getting Started guide for developer questions