Connect DeltaV events to batch review

Identify the source for each part of the record.

DeltaV’s historical products serve different evidence needs. Continuous history supplies selected process values; event history supplies alarms and events. A batch review also needs recipe execution context. Decide which source owns each required field before choosing a transfer protocol.

Source: Emerson: DeltaV Historians.

Verify the installed batch-history access route.

Emerson’s January 2017 Batch Historian data sheet describes recipe execution data from Batch Executive and process alarms/events from Event Chronicle. It also describes documented SQL views and the DeltaV History Web Service as external access routes. Use this as a route to investigate with the site owner, not a current-version installation instruction.

Obtain the installed DeltaV and historian versions, applicable licensed components, supported read documentation and a representative completed-batch extract. If documented SQL views are the approved route, request their schema and scoped read access. If a history service or completed export is used, capture its actual fields, filters and completion signals. Do not write a query against guessed table names.

Record the approved collector/gateway location and how it reaches the plant source. Prove that a read-only account retrieves the required historical population. OPC UA connectivity alone does not demonstrate that recipe hierarchy, batch events and historical measurements are all exposed by that endpoint.

Source: Emerson: Batch Historian product data sheet, January 2017.

Join on execution occurrence, not the word Hold.

Map source batch/run identity, recipe hierarchy, unit/equipment and the source phase execution identifier where available. In the exercise H-01 and H-02 are separate occurrences with the same displayed name. A join on batch plus phase name would merge their observations and hide the ten minutes between them.

If the export lacks an occurrence identifier, agree a deterministic derivation from the ordered source events and equipment context. Retain those source event identities and the derivation version. Overlapping or ambiguous starts, an end without a start and equipment reassignment need an unresolved outcome; assigning occurrence numbers from arrival order is unreliable.

Keep source event time, collector first-seen time and extract retrieval time as separate fields. The first 10:05 retrieval contains E-01 through E-05. The 10:10 retrieval contains E-01 through E-06. Match the five repeated event identities and add E-06 once; the two transfers represent six events, not eleven.

Reproduce elapsed time and active time separately.

H-02 spans 09:30–10:00: 30 elapsed minutes. The declared local rule excludes the five-minute pause, giving 25 active minutes. Keep both results and the interval list. The rule must be checked against the actual source state meanings; a Hold phase name is not itself a pause state.

An aborted, stopped or otherwise interrupted execution is not automatically a successful completion. Map the actual source terminal outcome and leave the review decision separate. If the source omits the end event, do not close the phase at collection time or assume a zero duration.

For a time-at-condition calculation, intersect the agreed execution intervals with usable measured-value intervals and the separately defined condition. An execution interval supplies context, not proof of temperature, pressure or agitation. Retain excluded pauses, quality gaps and unknown periods with their own durations.

A setpoint does not fill a missing observation.

For the second occurrence, suppose the recorded setpoint is 37.0 °C. A measured process value of 36.7 °C at 09:35 and another of 37.1 °C at 09:50 are observations at those times. Two points alone do not establish the temperature throughout the 25 active minutes. Preserve the actual measurement tag, equipment, units, source quality and interpolation or maximum-age rule.

If usable measurement coverage is absent from 09:50–09:55 under the agreed rule, that is five of the 25 active minutes: active-time coverage is 20 / 25 = 80%. Dividing by the 30-minute elapsed duration would answer a different question. Successful delivery of E-06 completes the event timeline but does not repair this measurement gap.

Keep setpoint, measured value and command/status fields distinguishable in Seal. Link the measurement extract to the occurrence record, with calculation inputs and revision. A reviewer can then see whether an exception is a process observation, an unavailable measurement, a missing event or an unfinished transfer.

Close a batch package only after its populations reconcile.

For each delivery, save source batch/occurrence selection, requested time bounds, extraction filters, retrieval time, event count, expected trend channels, continuation/completion evidence and File references. Compare the returned occurrences with the batch execution source. A closed last phase does not prove that earlier phases, alarms or selected trends were all retrieved.

Use an overlap or replay mechanism appropriate to the source to recover late events; a watermark based only on event time can miss E-06 if it arrives after that interval was declared processed. Reconcile repeated event identities and retained content. If the source provides a revision or changes a previously seen event, preserve that change as revised evidence rather than treating it as a new identical event.

A useful Seal setup links batch, occurrence, event extract and measurement extract, with typed durations and separate collection/review states. Begin with this two-occurrence package and exercise a network interruption. The receiving workflow is ready for the agreed scope when it recovers the missing evidence without duplicate occurrences and clearly retains the unresolved measurement interval.

Vendor references for this workflow

• Emerson: DeltaV Historians. The current product overview distinguishes continuous history, event history and viewing tools. It does not publish a universal batch-export schema.
• Emerson: Batch Historian product data sheet, January 2017. Pages 2–4 describe batch execution/event sources, documented SQL views and History Web Service access. This is an older specification: confirm installed-version support, licensing and the approved read route before implementation.

Test it with your data.

Use these cases to agree and test the connection’s behaviour. They are proposed acceptance checks, not completed tests or automatic connector features.

Check your system’s connection options.

Confirm the installed version, available exports and permissions. These pages cover the source interface and link to setup instructions.

• Emerson DeltaV
• AVEVA PI System
• OPC UA

For configuration: Seal Scripts and local-file collection.