Build forecast-validation data

Create labeled SWRO cases for challenging fouling forecasts

Track A · Synthetic data · TRL 3

Set the operating conditions and fault mix, then generate a labeled validation dataset here (one tidy CSV), or download the JSON spec — the same spec drives the Python generator (python3 export_dataset.py spec.json) for the full per-scenario file set.

Fault types
Severities
Negative controls
—
Dataset spec (JSON) 
{}

Schema aquasentinel/dataset-spec@1 — documented in BENCHMARK.md. The browser CSV is one long-format file (a scenario column per run); export_dataset.py spec.json writes the per-scenario files + manifest. The default suite includes the same eight named controls used by the locked protocol.