Real-Time Molten Pool, Arc & Stick-Out Analysis
Three process parameters analyzed simultaneously — molten pool geometry, arc stability, and wire stick-out — with AI defect scoring delivered in under 100 milliseconds.
Weld quality is determined in the first seconds of solidification. This demo shows the Therness AI engine tracking molten pool width, length, and aspect ratio frame-by-frame while simultaneously monitoring arc voltage anomalies and wire-stick-out deviation. All three signals feed a unified defect score — porosity risk, lack-of-fusion risk, and burn-through risk — logged against torch position for full-seam traceability.
Key Takeaways
- Simultaneous tracking of pool geometry, arc stability, and stick-out deviation.
- Sub-100 ms defect scoring — catches anomalies before the pool solidifies.
- Positional logging ties every defect event to seam coordinate for targeted rework.
- Works with GMAW, GTAW, FCAW, and robotic arc welding processes.
Deploy with Therness
- HeatCore™ thermal camera captures pool thermal signature; DuoSense AI™ adds arc monitoring.
- AI models train on your baseline process — no generic thresholds, no false positives.
- Alerts route to PLC for automatic speed/power correction or cell stop.
- ISO 17635 evidence package generated automatically per joint.
Transcript (short)
Show transcript
This video shows the Therness AI monitoring engine running live during arc welding. Three channels are analyzed simultaneously: molten pool geometry (width, length, aspect ratio), arc voltage stability, and wire stick-out distance. When any signal drifts outside the trained envelope, a defect-risk score is raised and the position is logged. The result is a per-seam quality record that supports ISO 17635 documentation without adding cycle time.
