Roughly every hour the plant's scenario engine wears out a tool on M010 and drifts the balancing head on M050. Below is that story reconstructed from the real archive: the scrap-rate ramp, the alarm at peak, the intervention, the recovery.
Fetching the last hours of telemetry.
After a stable phase, the engine ramps M010's defect rate (worn tool) and M050's residual imbalance over ~15 minutes.
Scrap climbs from a ~2% baseline toward ~20% at peak — the ramps you see in the chart, in real archived parts.
At peak, a CRITICAL alert fires on mes/alerts. For imbalance it points at M050 — even though the scrap shows up at M060's noise test.
An operator (or the AI assistant, or the auto-fix grace timer) triggers maintenance; the tool is changed and the rate falls back to baseline. Then the cycle re-arms.
Open the dashboard assistant during an alarm and ask "why is M060 scrapping?" — it will send you to M050.