A core-stacking sequence advances when the intended lamination is physically confirmed. Reconcile identity, detachment and receiving-layer state before committing the transaction.
Keep release and placement as separate states
A useful confirmation sequence is: tool at the defined placement condition; release commanded; physical detachment assessed; target-area presence assessed; position accepted; layer state updated. A failure or an inconclusive observation at any step enters a defined exception state. It does not silently become a successful placement.
Physical detachment can be checked through a suitable tool or part observation. Target-area presence needs evidence from the stack. A sheet can detach correctly and still miss the target, while a sheet partly supported by the tool can appear to occupy the target before withdrawal.
| State | Possible evidence | What it does not establish alone |
|---|---|---|
| Release commanded | Controller and valve record | Actual detachment |
| Tool clear | Tool-area observation after withdrawal | Presence at the intended layer |
| Stack response | Registered height or profile change | Correct identity and alignment |
| Placement accepted | Visible geometry matched to the expected part | Hidden sequence correctness without its separate record |
Combine observations with their blind spots
Height feedback can support presence detection, but the increment depends on the measurement location, contact state and stack compression. A tip placed outside the height-sensor field may leave the reading nearly unchanged. A contact gap can create an increment even without the expected full layer condition.
Vision can confirm exposed edges or a characteristic profile after tool withdrawal. It can also be fooled by an existing layer with the same outline or by an occluded area. Specify which features must become newly observable, and register the before and after views to the stack rather than to camera pixels alone.
Test missing-part scenarios directly
During controlled commissioning, challenge no sheet, delayed detachment, retained corner and off-target deposition. Include relevant interacting job/recovery states; NIST’s combinatorial-testing overview motivates interaction coverage, not proof of exhaustive testing [1]. Use controlled specimens and normal safeguards, recording the earliest detection and inconclusive outcomes.
Do not let a downstream height correction rewrite the earlier presence history. If the result is uncertain, preserve the expected layer, observed physical condition and exception disposition until the discrepancy is resolved.
For an assumed 1.0 m × 0.20 m sheet, catalogue thickness 0.23 mm and density 7,650 kg/m³ give mass 0.3519 kg. [2] Assume a post-release weighing witness has expanded uncertainty 0.010 kg. A zero-mass change and the expected 0.352 kg change are readily distinguishable, but a wrong equal-mass variant is not. Combine a physical presence check with the required geometry/identity evidence and a release-complete signal. The transaction closes only when the sheet is confirmed on the stack and absent from the gripper; a vacuum-pressure change alone can also represent a leak or dropped part.
Release timing, gripper occupancy sensing, receiving-stack support and part identity must be checked together. Retain challenge cases for a sheet left on the tool and one dropped beside the stack.
Record the expected part, confirming physical evidence and the event that advanced the layer state. This history exposes a missing lamination at the point where the sequence can be corrected.
References
- NIST, Combinatorial Testing, programme and method overview.
- Nippon Steel Corporation, Grain-Oriented Electrical Steel Sheets, catalogue D004je, April 2020, specification and typical-property tables.