H₂ Sensor Anomaly Detector

Gaseous Hydrogen Test Bench · 0–105 MPa · ML-Powered

Cycle 0

Speed

Model Comparison → ● CONNECTING

Test Pressure

— MPa

Gas Temperature

— °C

H₂ Flow

— ml/min

Specimen Temp.

— °C

Live Detection Model

Algorithm: Isolation Forest
Learning: Unsupervised
Library: scikit-learn 1.5
Training: 80 normal cycles
Contamination: 5 % (model param)
Features: p · T · flow · T_s · phase
Stream: 2 readings / sec
Medium: H₂ gaseous
Models: 5 compared →

0 readings

0 anomalies

— rate

Anomaly Log

No anomalies detected yet.

Dataset Export

Download a labelled CSV dataset for ML model comparison.

Cycles

↓ Download CSV