Processed train/validation/test splits for SMIB PINN vs ML benchmark from ANDES

This deposit provides processed trajectory data for supervised learning and evaluation on a single-machine infinite-bus (SMIB) benchmark used in an IEEE Access manuscript "Unified Physics-Informed Trajectory Surrogate for Transient Stability Assessment and Critical Clearing Time Search on a Single-Machine Infinite-Bus Benchmark" Contents (experiment id exp_20260211_190612):- train_data_20260211_190612.csv- val_data_20260211_190612.csv- test_data_20260211_190612.csv- all_splits_20260211_190612.csv (combined processed table for convenience / figure pipelines) Expected layout when used with the companion code repository: place the CSVs under: data/processed/exp_20260211_190612/relative to the repository root (see reproduction README in the code archive). Companion code (minimal public reproduction repository): https://github.com/albert8943/unified-physics-informed-neural-surrogate-smib-tsa-cct-benchmark Simulation context: trajectories were generated with the ANDES power system toolbox; the supervised learning stage consumes these fixed CSVs and does not require re-running the simulator for Path B–style reproduction when using this bundle alone. How to cite: cite this Zenodo record (DOI) together with the companion GitHub repository and the associated IEEE Access article once available.