Monitoring pump performance degradation

Hi ​@TIM 

the reason this never trips an alarm is that "the transfer is slow" isn't a threshold on any single tag; it's a per-event duration drifting upward. So the workflow is event-based, not threshold-based.

1. Turn each transfer into an event. VBS on pump running status (condition = Constant, or = running) with a duration cut → every transfer surfaces as one event.
2. Compute per-event KPIs in Event Analytics. Add calculations: event duration (your headline), plus avg motor current, avg flow rate, avg discharge pressure, Δ product level. You can attach up to 30.
3. See the deterioration. Scatter transfer duration vs. event start-time, a slow upward slope is your gradual degradation.
4. Find the signature with the parallel-coordinates plot. This is the key view here: each transfer is one line crossing axes for duration, current, flow, pressure and level. Brush the long-duration end of the duration axis and watch which other axes the highlighted lines pull toward that's your degradation fingerprint surfacing across all variables at once (e.g. long transfers tracking high current + low flow points at the pump end, not the line). Then drop to a scatter of duration vs. that one suspect to confirm and quantify the relationship. Parallel coordinates for the multi-KPI brush, scatter for the pairwise confirmation  that's the recommended refinement flow.
5. Persist and warn. Map these calculations to context-item fields so the history is a queryable record (mapping calcs to context fields), then once you know the signature, VBS on it and promote to a Monitor now you catch it before transfers become unacceptable.

If you want a full walkthrough of this event-based workflow, the Event Analytics in Action VILT recording covers detecting process events and comparing performance across runs end-to-end.

Kind regards
Frederik