In short. Monitoring replaces guesswork with evidence. Record voltage, current, frequency, harmonics, sags, swells and power factor over a representative period at the right points, then interpret the data against equipment tolerances.
What to measure
| Parameter | Why |
|---|---|
| Voltage per phase | Range, drift and imbalance |
| Current per phase and neutral | Loading and unbalance |
| Frequency | Grid and generator stability |
| Sags, swells and interruptions | Events that upset equipment |
| Harmonics and THD | Distortion sources and effects |
| Power factor and kVA | Capacity and penalties |
Where to place analysers
- At the incoming supply for the utility view.
- At the generator output for generator behaviour.
- At the sensitive load for what it actually sees.
- At large nonlinear loads for source identification.
Duration and settings
Record at least one full operating cycle, often a week, with event triggers set below equipment tolerance. Synchronise clocks so events line up across points.
Interpreting
Reading → baseline → trend → deviation → diagnosis → intervention → verification.
- Establish normal behaviour.
- Match events with plant operations.
- Decide which events are harmful.
- Choose the technology. See selection guide.
- Measure again after the fix.
Permanent monitoring
For critical sites, fixed meters with alarms and trends help catch drift before failure, similar to condition monitoring on generators. See preventive maintenance.
Illustrative example
A plant suffers unexplained drive trips. A week of data shows voltage dips at 6 pm each day, matching the start of a neighbour’s furnace, with a dip of about 12 percent for half a second. The fix is a ride-through on the drive supply, not a bigger breaker.
Defining measurement points in a drawing
Mark the monitoring points on the single-line diagram. A typical set: incoming supply, generator output, UPS input, UPS output, and the busbar of the most sensitive loads. This lets you compare what each stage does to the supply and find where a problem enters.
Choosing triggers
| Trigger | Typical setting idea |
|---|---|
| Sag | Below the lowest voltage the weakest load tolerates |
| Swell | Above the highest voltage the weakest load tolerates |
| Interruption | Near zero volts |
| Harmonics | Above agreed limits |
| Frequency | Outside the band set for the generator or process |
Interpreting a typical weekly report
- Look at the minimum and maximum voltage and at what time they occur.
- Compare with equipment tolerance.
- Count events and look for clusters.
- Check power factor and demand for billing.
- Examine THD and the harmonic spectrum.
- Report the findings in a one-page summary with a recommendation.
Comparison with standards
Measurement methods and classes are defined in recognised power-quality standards. Use analysers that comply with the relevant class for the purpose of the measurement, and state the method in your report.
Long-term monitoring value
Permanent meters on critical feeders create a history. A slow worsening of voltage or harmonics can be spotted before equipment fails. Combine with temperature monitoring for UPS rooms and stabilisers.
Data handling
Keep raw logs, the report and the decision. If a claim is later made against a supplier or the utility, dated records are the evidence.
The master engineering question
What does the data say happened in the moments before the equipment failed?
What to send us
Send logs or ask about monitoring. Send your application details.
Questions engineers ask
How long should I record?
A full operating cycle; often a week.
What accuracy is needed?
Use a recognised class-rated analyser.
Can I use the UPS display?
It helps, but dedicated analysers give more detail.
Should I monitor both utility and generator?
Yes, they behave differently.
Values are illustrative. Follow standards for method and limits.