Ensuring Data Quality in PI Vision through Effective Data Validation

PI Vision displays are only as reliable as the data and configuration behind them. A display can render correctly while the source value is stale, bad, incorrectly mapped, or no longer appropriate for the process.

Validation should therefore cover the full path from source to display.

Validate source data first

Check that important PI Points:

• Update at the expected rate

• Do not remain in bad states

• Use correct engineering units

• Show plausible process behavior

• Come from the intended source system

A display-level check cannot correct a bad source.

Validate AF references

Many PI Vision displays use AF attributes rather than direct PI Point references.

Confirm that the attribute points to the correct server, tag, and data reference. Recheck mappings after template changes, migrations, and equipment modifications.

A valid AF attribute can still point to the wrong signal.

Validate calculations

If a display uses calculated values, verify the analysis inputs, output, status, and schedule.

Check how the calculation handles missing or bad inputs. A calculation that continues to publish a numeric result during a source failure can hide the problem from the display user.

Check freshness at the display level

Operators need to know when a displayed value is not current.

For important displays, define the expected age of each signal and make stale conditions visible through symbols, status indicators, or supporting monitoring.

Do not use one freshness threshold for every measurement.

Review display dependencies after change

Tag renames, server changes, AF redesign, and equipment retirement can leave displays connected to obsolete objects.

Before a high-impact change, identify affected displays. After the change, verify both values and navigation paths.

Validate manual entries

If users enter values manually, apply range, data-type, timestamp, unit, and required-field checks before the value is written.

Record the user and reason for important corrections when the process requires traceability.

Prioritize critical displays

Not every display requires the same control.

Start with displays used for:

• Operations

• Reliability

• Production reporting

• Environmental reporting

• Critical maintenance decisions

Define an owner and a review cycle for these displays.

Use trust indicators carefully

A trust indicator can help users see that a source or calculation passed defined checks. The indicator must have a clear meaning.

For example, a trusted status can mean that the source is current, the data-quality checks passed, and the dependency configuration is known. It should not imply that the measurement is physically correct in every condition.

The objective

PI Vision validation should help users answer a simple question: can I use this value for this decision?

That requires data health, source traceability, valid calculations, and controlled display configuration, not only a display that loads successfully.