A validation result is a list of things to investigate, not a grade for the project. Start with one finding, open the record it names and work out whether the value is wrong or intentionally unusual.
Example: two I/O points share a tag
Suppose two rows both use LS-101. The finding identifies the affected records so you can compare them. If one was copied by mistake, correct that row. If the duplication has a valid engineering explanation, document the reason through the available exception process instead of inventing a value.
Work with current records
- Open the right project and finish any intended imports first.
- Have the relevant source list or drawing available. A warning alone does not tell you the correct engineering value.
- Make sure your role allows the correction or exception you intend to record.
Resolve one finding
- 01
Open Engineering validation and run the checks on the current project data.
- 02
Read a finding’s rule, severity and affected record. Start with a high-impact issue that you can investigate.
- 03
Open the named record and compare its data with the trusted source.
- 04
Correct the mistaken field and save. If it is an intentional exception, use the allowed acceptance action and enter a specific reason.
- 05
Run validation again. Check that the issue is resolved or has the documented exception status you intended.
Look at the source, not only the total
A lower warning count is useful only if the underlying records make sense. For LS-101, reopen both rows and confirm the final tags. A clean run covers the configured data rules; it does not certify the design or compare it with a live PLC.
If the finding comes back
- Check that you saved the actual record named in the finding, rather than a similar row.
- If the data changed after the previous run, investigate the latest result rather than treating the old report as current.
- A vague exception such as checked tells the next engineer very little. State the technical reason and the relevant reference.
Frequently asked questions
Must every warning be fixed by changing data?
No. Some unusual values are intentional. An authorized, explained exception can be more honest than a fabricated correction.
Can I resolve findings as a viewer?
A view-only role can inspect available results but cannot make the project changes needed to resolve them.
Will an import remove earlier findings automatically?
Treat the imported data as a change and run the checks again to obtain a result for the new state.
Should I wait until handover to run validation?
No. Smaller, regular reviews make it easier to trace a problem to the change that introduced it.