DIII-D SOLPS-ITER: Deuterium Plasma Simulations for Neural Surrogate Training

SOLPS-ITER steady-state simulations for a deuterium plasma in the DIII-D tokamak, prepared for machine-learning applications. The 16.7 GB dataset was created by Abdourahmane Diaw and last updated on April 17, 2026. Simulations use a validated lower-single-null configuration based on DIII-D discharge 174310.

Use Cases

• Training neural network surrogates based on validated plasma simulation outputs.
• Analyzing steady-state divertor conditions based on the described lower-single-null configuration.
• Studying neutral particle dynamics in fusion plasmas based on the referenced validation.
• Benchmarking machine learning models for plasma physics based on the simulation dataset.

Strengths

• Dataset size is 16.7 GB, indicating a substantial volume of simulation data.
• Simulations are validated against an experimental DIII-D discharge, as referenced in the description.
• Released under a CC-BY-4.0 license, permitting open use and modification.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for specific ML tasks.

Provenance

Source

Author Abdourahmane Diaw, shared via figshare.

Collection Method

Collection of steady-state SOLPS-ITER plasma simulations.

Time Range

Simulation time point validated against DIII-D discharge 174310 at t = 3500 ms.

Freshness

Last updated 2026-04-17 22:52:33; freshness should be verified.

Geography

Data is from the DIII-D tokamak experimental facility.