Level Data Approval for Multiple Systems: Tags, Quality and Ownership

When one level measurement feeds a PLC, DCS, historian, dashboard and business report, each copy can acquire a different name, unit, range, timestamp or quality rule. The field transmitter may be correct while one consumer shows percent, another uses an old tank table and a third holds stale data without warning. Site approval should define a single measurement meaning and controlled transformations, then prove that every system preserves identity, value, status and time. Ownership is as important as protocol configuration because later changes must remain coordinated.

Define the authoritative measurement

Choose the authoritative raw value and document whether it is distance, level, percent, volume or mass. State reference datum, range, units, resolution, update interval and quality. Identify where each conversion occurs and who owns it. Retain raw level when possible so tank-table or density changes do not destroy traceability.

Create a tag and interface map

Map instrument tag, channel, register or object, PLC tag, DCS tag, historian point, gateway topic and cloud field. Include data type, byte order, scaling and unit. Use consistent names without assuming every platform accepts the same length or characters. The map should expose duplicate calculations and orphaned tags before commissioning.

Preserve quality and age

Carry device fault, lost echo, over-range, maintenance, communication loss and data age through each interface. Define stale thresholds based on expected update rate and operational need. Do not replace invalid values with zero or retain the last value as healthy. Dashboards and reports should show or filter quality instead of hiding it from users.

Control timestamps and time zones

Choose whether timestamps originate at the instrument, RTU, control system or platform and synchronize clocks. Preserve original measurement time during buffering and backfill. Store time zones or UTC consistently and handle daylight-saving rules where relevant. Test delayed and out-of-order data so historical trends and alarms do not misrepresent the event sequence.

Govern scaling and tank tables

Assign one controlled owner for offsets, linearization, volume tables and density compensation. Version every table and record its effective date. Avoid scaling in both transmitter and PLC unless the architecture explicitly requires it. Compare several independently calculated points and protect historical data from silent reinterpretation after a table revision.

Separate monitoring from control

Document which systems may only read and which may write parameters or issue control commands. Restrict access by role and record changes. A cloud or business dashboard should not become an undocumented configuration path to a field instrument. Critical control and protection should retain the availability, independence and proof testing defined by the site risk assessment.

Test all consumers and failures

At known low, middle, high and alarm points, compare field value, raw interface, every system tag, display and report. Test sensor fault, bad quality, communication interruption, stale data, power recovery and backfill. Verify alarm recipients and final actions. Remove forces and confirm each consumer resumes the authoritative live source.

Maintain a cross-system change record

Record reason, owner, approver, time and affected tags for range, unit, table, filter, alarm, protocol or software changes. Keep backups aligned and schedule regression checks for critical consumers. Retain the interface map and acceptance evidence under the measurement tag. This prevents a local software change from silently altering operational or commercial meaning elsewhere.

Engineering checklist

• Define one authoritative raw measurement.
• Map tags, types, units and conversions.
• Preserve quality and measurement age.
• Control timestamps and table revisions.
• Restrict write access and separate control.
• Test every consumer during normal and fault states.

Frequently asked questions

Which system should calculate volume?

Use one controlled location with a named owner and revisioned tank table; retain raw level for traceability.

Can a dashboard show the last value during an outage?

Yes, but it must clearly identify the value as stale with its original timestamp.

What requires regression testing?

Changes to scaling, units, tables, alarms, protocols or tag mappings should be checked across every dependent consumer.

Related measurement solutions

80 GHz radar level meters · Industrial IoT gateways