Engineering QA / Evidence and digital-twin calibration

Evidence before approval.

Compare synthetic submitted values and inspect a quality gate without claiming actual SITL execution, real sensor calibration or production release. Metrics and record checks are independently important.

Synthetic timeseries review

Not measured

Metric name

Synthetic prediction bias: 0.20 m

Alternating synthetic deviation: 0.10 m

RMSE (m)

0.233

MAE (m)

0.211

Mean bias (m)

0.211

Error arithmetic is valid for the submitted numerical series, but the numbers are invented. It does not verify an aircraft, digital twin or physical test.

Manufacturing release evidence

Missing evidence

Training-only checks, not genuine signed records. Mark each category as hypothetically documented.

Engineering BOM revision Manufacturing BOM revision Supplier lot genealogy Inspections complete Firmware hash evidence Independent QA reviewer identified

Open major nonconformances

Missing categories6

Production releasedNO

ebom: not documented
mbom: not documented
lots: not documented
inspection: not documented
firmware: not documented
reviewer: not documented

Even a complete training checklist cannot sign a real manufacturing release, certify an aircraft or grant permission to fly.