ConstellaRation
State-of-the-art dataset for fusion energy research, with stellarator geometries and MHD metrics, ready to use for AI-assisted optimization.
Description
ConstellaRation is a scientific dataset designed to facilitate research on stellarators, magnetic confinement devices for fusion. It includes plasma border shapes, ideal magnetohydrodynamic (MHD) metrics, and associated sampling parameters. The game is structured in two tabular parts interconnected via identifiers, with a format ready for VMEC2000 simulations.
What is this dataset for?
- Optimizing stellarator design through machine learning
- Evaluate the stability and performance of simulated plasma geometries
- Accelerating nuclear fusion research through MHD simulation
Can it be enriched or improved?
Yes. It is possible to cross this dataset with other plasma simulations, to add annotations on energy performance or technical feasibility, or to explore geometric variations via generative algorithms. It is also possible to automate stability analysis via predictive models trained on this data.
🔎 In summary
🧠 Recommended for
- Nuclear fusion researchers
- MHD simulation engineers
- Geometric optimization specialists
🔧 Compatible tools
- VMEC2000
- Python (Pandas, NumPy)
- Scikit-learn
- TensorFlow
- PyTorch
💡 Tip
Use the plasma_config_id column to relate geometries to MHD simulation results and facilitate comparative analysis.
Frequently Asked Questions
Does this dataset make it possible to generate geometries for VMEC simulation?
Yes, JSON files can be converted into VMEC2000 files that can be used for MHD stellarator simulation.
Is it suitable for use in machine learning?
Yes, data is structured to be used in optimization, regression, or even generative learning.
Can it be used without advanced knowledge in physics?
This dataset is primarily intended for fusion experts, but collaborations with data scientists can make it accessible to other profiles.
