Power quality is a maintenance signal, not just a compliance topic

Power quality usually enters the conversation after a trip or a utility dispute. That is late. The same waveform data that documents an event also describes the slow degradation of the equipment attached to it — which means a power quality program that only produces compliance reports is leaving most of its value on the table.

The reframe is straightforward: treat the electrical supply as a condition-monitoring channel. Vibration tells you about bearings. Thermography tells you about connections. Power quality tells you about the electrical stress a machine is absorbing every hour it runs, and that stress is a leading indicator of when it will fail.

Why events are only half the story

Event capture is the familiar use case. A sag drops a line, the system timestamps it, and you have evidence for the utility conversation. Useful, but reactive by definition.

The predictive half lives in the trends between events. Equipment rarely fails because of one bad moment. It fails because of thousands of unremarkable ones — a motor running slightly imbalanced for two years, a transformer carrying harmonic current it was never derated for, a drive absorbing repeated shallow sags that each cost a little insulation life.

None of those produce an alarm. All of them produce a trend.

What to trend continuously

A handful of metrics carry most of the predictive value.

• Total harmonic distortion against IEEE 519 limits, per feeder. Track voltage and current distortion separately. Current THD rising on a feeder with no load change usually means something downstream has degraded.
• Voltage imbalance on three-phase motor circuits. The rule of thumb worth remembering: a small percentage of voltage imbalance produces a much larger percentage of current imbalance, and the extra current turns into heat in the windings. Motors on chronically imbalanced supply run hotter than their nameplate assumes and age accordingly.
• Sag frequency and depth, mapped to affected equipment. Count them per circuit and per month. A rising count is a system telling you something about your supply, your loads, or both.
• Neutral current on shared circuits. Elevated neutral current in a system with heavy single-phase non-linear load is a fire and overheating risk long before it is a metering curiosity.
• Power factor by feeder, not just at the service. Site-level power factor can look acceptable while one feeder is dragging badly. The penalty is site-level; the fix is feeder-level, and you cannot find it without the granularity.

Reading the trends as maintenance intelligence

Each of these metrics maps to a physical failure mode, which is what makes them actionable rather than merely interesting.

Rising current distortion on a fixed load typically points to a drive with a degrading DC bus capacitor or filter. It is one of the few electrical signatures that gives months of warning.

Growing voltage imbalance often traces to a loose or corroding connection, a failing phase on a bank, or a single-phase load added to one leg without rebalancing. All three are cheap to fix when found and expensive after a motor cooks.

Increasing sag sensitivity — the same depth of sag now dropping equipment that used to ride through — suggests a control power supply losing headroom rather than a worsening grid.

Creeping neutral current points to an accumulating population of non-linear single-phase loads, which is a design conversation before it is a maintenance one.

Turn events into evidence

When an event does cause downtime, the timestamped record is what makes a utility conversation productive and what tells maintenance which asset absorbed the hit.

Evidence has to be assembled before you need it. That means:

• Monitoring at the point of common coupling and at the affected equipment, so you can distinguish a supply problem from one you generated internally.
• Retaining enough interval and waveform history to establish a pattern, not just the last incident.
• Recording the downstream consequence — which line stopped, for how long, at what production cost. A sag record with no cost attached is a technical footnote; the same record with a dollar figure funds the mitigation.

That last point decides most mitigation budgets. Ride-through equipment, line reactors and correction gear are all straightforward purchases when the loss they prevent is documented, and impossible to justify when it is anecdotal.

Where to start

You do not need waveform capture everywhere on day one. Start where interruption is most expensive and where motor load is concentrated:

1. Instrument the main service for the utility-facing picture and penalty exposure.
2. Instrument the two or three feeders serving critical or motor-heavy loads.
3. Establish a baseline over a normal month before drawing any conclusions.
4. Set alerts on rate of change, not just on limits. A THD figure inside IEEE 519 that has doubled in six months is more informative than one sitting steadily near the limit.
5. Route those alerts to maintenance, not just to the energy report.

Step five is the one most programs skip. Power quality data that lands only in a compliance folder cannot change a work order, and changing work orders is where the return comes from.

The point

Compliance is a floor. The same instrumentation that keeps you inside the limits and settles utility disputes is, with no additional hardware, a continuous health monitor for every motor, drive and transformer on the site. Treating it that way turns a reporting obligation into a maintenance advantage.