8000
Skip to content

keisuke58/Payload_gnn

Folders and files

NameName
Last commit message
Last commit date

Latest commit

ย 

History

298 Commits
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 

JAXA H3 PyTorch Geometric Abaqus License

๐Ÿš€ GNN-SHM: Graph Neural Networks for H3 Rocket Fairing Structural Health Monitoring

Debonding detection on CFRP/Al-Honeycomb sandwich structures using Geometry-Aware GNNs

Quick Start โ€ข Features โ€ข Pipeline โ€ข Citation โ€ข ๐Ÿ“š Wiki โ€ข Contributing


๐ŸŒŸ Overview

GNN-SHM is a research project that combines Graph Neural Networks (GNN) with Finite Element Method (FEM) to detect and localize skin-core debonding in the JAXA H3 rocket's CFRP/Aluminum Honeycomb payload fairing.

ๆ—ฅๆœฌ่ชž: JAXA H3 ใƒญใ‚ฑใƒƒใƒˆใฎ CFRP ใƒใƒ‹ใ‚ซใƒ ใ‚ตใƒณใƒ‰ใ‚คใƒƒใƒใƒ•ใ‚งใ‚ขใƒชใƒณใ‚ฐใซใŠใ„ใฆใ€GNN ใจ FEM ใ‚’็ตฑๅˆใ—ใŸใ‚นใ‚ญใƒณ-ใ‚ณใ‚ข็•Œ้ขใƒ‡ใƒœใƒณใƒ‡ใ‚ฃใƒณใ‚ฐไฝ็ฝฎ็‰นๅฎšใ‚ทใ‚นใƒ†ใƒ ใ‚’้–‹็™บใ€‚2025ๅนด F8 ไบ‹ๆ•…ใง้ก•ๅœจๅŒ–ใ—ใŸ CFRP/Al-HC ๆŽฅ็€ๅฅๅ…จๆ€งใƒขใƒ‹ใ‚ฟใƒชใƒณใ‚ฐใฎๅฎŸ็”จๅŒ–ใ‚’็›ฎๆŒ‡ใ™ใ€‚

Why This Matters

H-IIA/B, Epsilon (Legacy) H3 (This Project)
Skin Al 7075 CFRP (T1000, AFP)
CTE Mismatch โ‰ˆ0 Severe (โˆ’0.3 vs 23 ร—10โปโถ/ยฐC)
SHM Need Low (40yr mature) High (F8 accident, 2025)

The H3 F8 accident (Dec 2025) identified CFRP/Al-Honeycomb interface debonding as a likely cause. This project aims to enable Condition-Based Maintenance (CBM) via guided-wave SHM + GNN-based defect localization.


๐ŸŽฅ H3 Rocket Launch (JAXA)

H3 Rocket Test Flight No.2 Lift-off (Source: JAXA Digital Archives)


โœจ Features

  • Geometry-Aware Graph Construction: Surface normals, principal curvature, geodesic distance โ€” no UV-mapping distortion
  • 4 GNN Architectures: GCN, GAT, GIN, GraphSAGE with Focal Loss for class imbalance
  • H3-Spec FEM: Barrel + Ogive (ฯ†5.2m), thermal load (CTE mismatch), debonding defects
  • Cutting-Edge ML Roadmap: Graph Mamba, E(3)-Equivariant GNN, FNO surrogate, PINN
  • Multi-Class Target: debond / delam / impact / healthy (2-year roadmap)
  • JAXA Collaboration: Real PSS test data validation planned

๐Ÿ— Pipeline

flowchart LR
    subgraph DataGen["๐Ÿ“ฆ Data Generation"]
        DOE[generate_doe.py<br/>DOE params]
        BATCH[run_batch.py<br/>Abaqus FEM]
        DOE --> BATCH
    end

    subgraph FEM["๐Ÿ”ฌ FEM"]
        ABAQUS[Abaqus FEM<br/>H3 Fairing<br/>Thermal + 120ยฐC<br/>Debonding]
    end

    subgraph Graph["๐Ÿ“Š Graph"]
        EXTRACT[extract_odb<br/>CSV]
        BUILD[prepare_ml_data<br/>Curvature-Aware Graph]
        EXTRACT --> BUILD
    end

    subgraph Train["๐Ÿง  GNN Training"]
        GNN[GCN / GAT / GIN / SAGE<br/>Focal Loss]
    end

    subgraph Deploy["๐Ÿš€ Inference"]
        INFER[Checkpoint]
        HEATMAP[Defect Prob<br/>Heatmap]
        API[FastAPI<br/>REST API]
        INFER --> HEATMAP --> API
    end

    BATCH --> ABAQUS
    ABAQUS --> EXTRACT
    BUILD --> GNN
    GNN --> INFER
Loading

๐Ÿš€ Quick Start

# Clone
git clone https://github.com/keisuke58/Payload_gnn.git
cd Payload_gnn

# Install
pip install -r requirements.txt

# Train (existing data)
python src/train.py --arch gat --epochs 200 --cross_val 5

# Inference API
MODEL_CHECKPOINT=runs/<run>/best_model.pt uvicorn src.predict_api:app --port 8000

Full Pipeline (with Abaqus)

python src/generate_doe.py --n_samples 50 --output doe.json
python src/run_batch.py --doe doe.json --output_dir dataset_output
python src/build_graph.py --data_dir dataset_output
python src/train.py --arch gat --epochs 200

๐Ÿ“ Project Structure

Payload2026/
โ”œโ”€โ”€ src/                    # Core pipeline
โ”‚   โ”œโ”€โ”€ generate_fairing_dataset.py   # Abaqus FEM
โ”‚   โ”œโ”€โ”€ build_graph.py               # Curvature-aware graph
โ”‚   โ”œโ”€โ”€ train.py                     # GNN training
โ”‚   โ””โ”€โ”€ predict_api.py              # FastAPI inference
โ”œโ”€โ”€ wiki_repo/              # ๐Ÿ“š Full documentation
โ”‚   โ”œโ”€โ”€ Home.md             # Wiki index
โ”‚   โ”œโ”€โ”€ Cutting-Edge-ML.md  # Graph Mamba, Equivariant GNN
โ”‚   โ”œโ”€โ”€ Vocabulary.md       # Technical glossary
โ”‚   โ””โ”€โ”€ ...
โ”œโ”€โ”€ .github/
โ”‚   โ”œโ”€โ”€ ISSUES.md           # Task index
โ”‚   โ””โ”€โ”€ ...
โ””โ”€โ”€ requirements.txt

๐Ÿ“Š Dataset

Item Value
Graphs 101 (train 81 + val 20)
Nodes/graph ~10,897
Node features 16 (normal, curvature, stress, temp)
Edge features 5

Defect size distribution: Small 30%, Medium 40%, Large 25%, Critical 5%.

Dataset Progress Visualization

**โ†’ Full visualization (docs/DATASET_VISUALIZATION.md)

Summary Feature Distributions
Summary Features

๐Ÿ“š Documentation

Resource Description
Wiki Home Full project index, status, navigation
2-Year Goals 5K samples, 4-class, Sim-to-Real
Cutting-Edge ML Graph Mamba, Equivariant GNN, FNO, PINN
Vocabulary Technical terms (ENโ†”JP)
Publication Venues IWSHM, JSASS, Structural Health Monitoring journal

๐Ÿค Contributing

We welcome contributions! See CONTRIBUTING.md for guidelines.


๐Ÿ“„ Citation

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

@software{payload_gnn_2026,
  title = {GNN-SHM: Graph Neural Networks for H3 Rocket Fairing Structural Health Monitoring},
  author = {Payload2026 Contributors},
  year = {2026},
  url = {https://github.com/keisuke58/Payload_gnn}
}

๐Ÿ“œ License

This project is licensed under the MIT License - see LICENSE for details.


๐Ÿ™ Acknowledgments

  • JAXA โ€” H3 specifications, collaboration
  • Open Guided Waves โ€” Benchmark dataset
  • PyTorch Geometric โ€” GNN framework

If this project helps your research, please consider giving it a โญ

About

GNN-based Structural Health Monitoring for CFRP/Al-Honeycomb rocket fairing โ€” debonding defect detection using Graph Neural Networks with FEM-generated datasets

Topics

Resources

License

Code of conduct

Contributing

Security policy

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

0