CASE 02 · CASE ATLAS · LEVEL B WORKSHOP SCENARIO

Blade bonding cure conditions

SITUATION

A blade-bonding line has an intermittent void rate that inspection has not yet linked to a specific cure parameter. Process engineering wants a preventive signal before the next non-conformity, not a retrospective root-cause report.

EVIDENCE

Cure-cycle logs (temperature, pressure, dwell time) and the void-rate inspection record for each cured part, joined by lot and station.

AI APPROACH

A statistical or anomaly-detection method — not a language model — flags cure-cycle windows that correlate with elevated void rate, ranked by strength of association, for a process engineer to investigate as candidate causes.

KEY CONTROLS

The output is explicitly a candidate-correlation list, never a stated cause; the tool's output destination is fixed as an investigation input, not an automatic process-parameter change.

VALIDATION MEASURES

Compare flagged windows against process engineering's independent judgement on a held-out set of prior lots; track false-flag rate so investigation effort is not wasted chasing noise.

HUMAN DECISION

Process engineering opens or declines a preventive investigation based on the flagged correlation and owns any resulting parameter change through normal change control.

DISCUSSION QUESTIONS

  1. What would change if this tool's output silently started adjusting cure parameters instead of only flagging them?
  2. How do you tell a real preventive signal from a coincidental correlation in a small lot sample?
  3. Who re-baselines the correlation model when the bonding recipe changes?

TRY IT LIVE

Run this scenario's approach through Data & Evidence Readiness — Is the evidence foundation ready for a controlled AI use case?

Open Data & Evidence Readiness →