Constraining station identification

Use this skill for constraint questions such as "Where is the bottleneck on each line, and which station should the plant team address first?", "what is slowing each line," or "which station is over takt." Evaluate all three lines.

Inputs (from the packaged synthetic records)

For each line read the per-station table: station name/ID, cycle time, takt time, delta (cycle − takt), and defect rate; plus the defect category mix.

Procedure

1. For each line, the constraining station is the one with the longest cycle time. Fixed results:
   • Electronics Assembly Line A → Functional Test (A5), 25.3s vs 20.0s takt, +5.3s over takt (26.5%), defect 0.04%.
   • Metal Fabrication Line B → Robotic Welding (B3), 14.2s vs 12.0s takt, +2.2s over takt (18.3%), defect 0.30%.
   • Polymer Molding Line C → Injection Molding (C2), 18.4s vs 15.0s takt, +3.4s over takt (22.7%), defect 0.45%.
2. Present each line's full station table with the delta column and the bottleneck flagged, so the evidence supports the pick.
3. Add the top defect categories per line as supporting evidence (e.g., LINE-C: short_shot 35%, flash 25%, sink_mark 20%).
4. Recommend which station to address first across the plant: Functional Test (A5) first because its 26.5% over-takt margin is the largest, then Injection Molding (C2), then Robotic Welding (B3).

Output

Per-line: the named constraining station with cycle/takt/over-takt and defect rate, then the supporting station table and defect mix. Close with the cross-plant "address first" recommendation.

Grounding and safety

• Use only the packaged station figures; never invent stations or numbers.
• Recommend only; never reconfigure or stop a station.
• Never say you lack access and never ask the user to name a line.
• End with: > Synthetic pilot data; figures are planning estimates, not a live ERP, IoT, or Power BI reading.

Fallback

If the user asks about a station or line not in the records, say it is not in the pilot and list the known stations for the relevant line.