A nuisance trip is reset. A UPS alarm is acknowledged. A panel is checked and feels warm but not alarming. A brief production interruption gets written up as a one-off.
Any of those, once, is a maintenance task. The difficulty is that electrical infrastructure rarely goes from stable to failed without leaving a trail first, and the early entries in that trail look exactly like the small, forgettable events above.
The question worth asking, every time, is not what failed but whether this is the first time.
Why the small symptoms matter
Deterioration in electrical systems is usually gradual and often intermittent. A breaker operates unexpectedly. A UPS alarm clears itself before anyone reaches the panel. Equipment resets with no obvious trigger.
Individually, none of it justifies an investigation. Repeated over months, the same events can indicate insulation degradation, thermal stress, load imbalance, degraded connections or a power quality problem sitting upstream of everything you’ve been replacing.
There’s good evidence for how much warning this trail actually offers. A study of live partial discharge monitoring on paper-insulated 11 kV cables in London, found that 12 of 17 cable sections showed clearly rising PD activity between 5 and 150 days before they failed. The signal was there, well ahead, in nearly three quarters of cases.
Whether anyone was reading it is a separate question.
Common indicators and what sits behind them
The same visible symptom can have several possible causes, and they often sit at different points in the system. The table below is a starting point for narrowing it down.
| What you’re seeing | Possible local cause | Possible wider cause |
|---|---|---|
| Nuisance tripping / repeated breaker operation | Overload, local equipment fault, poor protection discrimination | Harmonic distortion, transient events, insulation deterioration, upstream supply quality |
| Overheating components or thermal hotspots | Loose or high-resistance connection, overloaded circuit, poor ventilation | Harmonic currents, phase imbalance, sustained load growth beyond design |
| UPS alarms or instability | Battery or module fault within the UPS | Unstable input voltage, frequency variation, voltage dips, poor incoming supply quality |
| Unexplained equipment resets, PLC or control faults | Control wiring, local supply fault | Voltage dips from large motor starts or network events |
| VFD / drive trips | Drive fault, motor fault, mechanical load | Voltage instability, harmonics, switching events, load changes |
| Flickering lights | Local circuit loading | Voltage fluctuation, flicker from cyclic loads on site |
| Transformer noise or temperature rise | Loose core clamps, cooling issue | Harmonic loading, DC offset, sustained overload, phase imbalance |
| Contactor dropout | Coil or contact wear | Short-duration voltage dips below the coil’s hold-in threshold |
| Burning smell, discoloured terminations | Loose connection, arcing, moisture ingress | Cyclic thermal stress from load or harmonic current |
| Power factor charges appearing on bills | New inductive load | Changed load profile, correction equipment failed or switched out |
The visible symptom is rarely the root cause
Two examples are worth spelling out with numbers, because they explain why competent investigations reach wrong conclusions.
A voltage dip you never recorded can trip equipment that appears faultless. A peer-reviewed IET review of the GB distribution system compiled equipment sensitivity data from IEEE research: AC-coil motor contactors can drop out at between 30% and 75% of nominal voltage, in as little as 10 to 80 milliseconds. Adjustable speed drives are affected in the 60% to 85% range over 10 to 170 ms. PLCs vary more widely still.
A dip of that depth and duration is over before anyone notices the lights. The contactor drops, the process stops, and the investigation finds a contactor that tests perfectly, because it is perfect. It did what it was designed to do at a voltage nobody measured.
Overheating is frequently an upstream problem manifesting locally. Manufacturer data for LV air circuit breakers shows that at 40% current harmonic distortion, heat losses run around 10% higher, the thermal equivalent of pushing roughly 5% more current through the device than the design intended. Replace the hot component and the replacement will run hot too.
Why the timestamp matters more than the fault description
When a fault repeats, the most valuable piece of information is what else was happening at that moment?
- Was the site at peak demand?
- Did it coincide with a specific machine starting or a production changeover?
- Were multiple assets affected in the same second?
- Was there a voltage or current disturbance at the same instant?
Without that context, investigation time goes into the wrong hypothesis.
A reset or a replacement removes the symptom, and if the underlying condition is unchanged, the fault returns on its own schedule.
There’s also a reporting gap worth knowing about.
GB supply reliability is measured through CI and CML indices, which, as the same IET review notes, only count interruptions longer than three minutes. Every dip, transient and momentary interruption falls below that threshold. The review also found that short interruptions in GB rose by 16% between 2010 and 2015 even as the headline reliability figures improved.
So the reliability data your DNO reports can be genuinely good while the disturbances that actually stop your plant are increasing, unmeasured, underneath it.
Moving from “it keeps happening” → “this is why”
- Monitoring may show that equipment resets coincide, to the second, with voltage dips on the incoming supply.
- Thermal data may reveal a hotspot developing over weeks.
- Power quality analysis may identify the harmonic source.
- Load data may expose an imbalanced circuit.
- Partial discharge monitoring may indicate insulation deterioration inside an HV asset that looks entirely normal from outside.
That’s the shift from “we keep seeing the same problem” to “we know what’s causing it and what needs to happen.”
When is further investigation justified?
- the same fault keeps returning after being fixed
- several assets are affected at the same time
- equipment issues coincide with voltage events
- components repeatedly overheat after being replaced
- faults started after new machinery, PV, battery storage or EV charging was commissioned
- a temporary repair has already failed once
Any two of those together usually means the cause sits somewhere other than where the symptom is appearing.
If this is already happening on your site...
Recurring faults that survive competent investigation are usually an evidence problem rather than an engineering one.
The measurement needed to identify the cause wasn’t in place when the event occurred.
Acteniq investigates recurring and unexplained electrical faults using power quality analysis, event capture, condition monitoring and partial discharge diagnostics, then delivers the corrective work and verifies it against the baseline. We look at the system, not the device.
Read how we traced recurring UPS alarms and machine stoppages back to their actual cause
Dealing with a fault that keeps coming back?
Tell us what you’re seeing and we’ll tell you what evidence would identify it.
Sources cited in this piece
- Walton, “Detecting and locating MV failure before it occurs: experience with live line partial discharge detection on underground paper insulated 11 kV cables in London”, CIRED 2001, via Zhang et al., “Review on Detection and Analysis of Partial Discharge along Power Cables”, Energies 14:7692, 2021 (open access)
- Vegunta, Watts, Milanović, Djokic & Higginson, IET Generation, Transmission & Distribution, 2019: equipment sag tolerance (Table 1, sourced to Djokic et al., IEEE Transactions on Power Delivery, 2004 to 2005), CI/CML three-minute threshold, and the 16% rise in short interruptions
- Schneider Electric, Masterpact NT and NW Maintenance Guide, LVPED508016EN-02, 07/2013: harmonic heat loss
