Our FR10 plate-check monitor adds a completion when it reconnects to a PLC bit already high. I understand why that's a duplicate if the result was saved earlier. What if the result happened while the monitor was away?
Someone suggested starting with the current bit value and waiting for the next change. I'm not sure that handles both cases.
It doesn't account for the missing history. Compare the retained job/result identity with durable application records. If that link is unavailable, show uncertainty instead of inventing a new event.
We store a plate description, not a unique job reference. The same plate is used repeatedly, so matching that description would collapse legitimate checks too.
Exactly. Part identity and job identity answer different questions. Your interface needs enough retained information to distinguish repeated checks of that plate, including restart boundaries.
Thanks, that distinction made the review clearer. We've kept reconnect as an unresolved bookkeeping state until the job-reference change is agreed. No automatic count increment and no automatic repeat request.
Give that unresolved state an owner and an explanation on screen. Otherwise someone will interpret the frozen total as another connection fault and keep reconnecting.