The majority of plants do not lose production due to catastrophic failures. They lose it in three-second sags nobody sees coming, tripped contactors, and controllers that reboot mid-batch. The power quality is not considered background noise; heavy industry nowadays treats it as a maintenance discipline with its diagnostics, budget, and workforce requirements. Or else, you will end up paying for the damage post fact.
What Actually Goes Wrong On The Line
Power fluctuations differ from one another. Voltage sags (dips) are brief reductions in the voltage supply, generally due to a disturbance on the grid somewhere upstream – a lightning hit on a transmission line, a large motor starting somewhere else on the network, or a fault that is cleared by a breaker three substations away. Swells are short increases in voltage. Transients and voltage spikes are even sharper and shorter, often triggered by a switching function or capacitor bank re-energizing, and contain enough energy to penetrate insulation that is already weakened.
Harmonic distortion is not a fluctuation in the typical sense but acts like one at the equipment level – current waveforms that are distorted and cause transformers and neutral conductors to overheat even when the voltage looks normal on a multimeter.
Of all of the above, voltage sags due to external grid faults are the most common and the most costly. They are not spectacular. They are hardly ever seen as an alarm. But they make up the most of production-related power quality complaints in facilities that run continuous processes as a sag doesn’t have to last for long to cause damage.
Why Sags Cost More Than Anyone Budgets For
A voltage sag of three cycles can drop a motor contactor on a motor control center, and once that contactor drops, the motor doesn’t just slow down – it stops, and the process behind it stops with it. Digital control systems reset. PLCs lose state. Drives fault out on undervoltage protection and won’t restart without manual reset, sometimes with a full sequence check before anyone’s willing to bring the line back up.
A sustained overvoltage is a more insidious villain. It also accelerates insulation aging in motors and transformers, shortening equipment life in ways that don’t show up until years later as a bearing failure or a winding fault that looked, on the surface, unrelated to power quality at all. Transients are the sharpest threat – short-duration, high-magnitude events that stress insulation directly and can take out sensitive electronics in a single event.
The thing is, this isn’t a small problem. A widely-cited EPRI study estimated that power outages and power quality disturbances cost industrial and digital economy businesses somewhere between $119 billion and $188 billion a year. The study’s old, but the loss mechanism is the same – plants absorb these costs quietly, spread across scrap, downtime, and equipment replaced early, rather than as a single line item anyone can point to.
Measure Before You Spend
The instinct after a bad month is to buy protection. That’s usually the wrong first move. Mitigation without measurement tends to target the wrong disturbance type – a facility installs surge protection when its real problem is sustained harmonic distortion from a bank of VFDs, or it buys a UPS sized for momentary sags when the actual event lasted four seconds.
A proper power quality audit uses PQ analyzers at the service entrance and at the sensitive load itself, run over enough time to capture the actual disturbance profile rather than a snapshot. This tells you whether you’re dealing with sags, transients, harmonics, or some combination, and it tells you the magnitude and duration you actually need to protect against. Skipping this step is how plants end up with expensive equipment sitting next to the exact fault it was never designed to solve.
Building The Mitigation Stack
Once you identified the problem, the solution is not just one product. It is a combination of approaches to mitigate the risk. Earthing and surge protection are at the bottom – the former doesn’t work without proper grounding, and the same grid that prevents blown fuses in the electrical room saves human lives when a shovel contacts a power cable.
On top of that, you have power conditioning for the least robust equipment, in terms of voltage tolerance: UPS for PLCs and PC-based control systems, voltage regulation (ferroresonant transformers, or automatic voltage regulators) for equipment sensitive to voltage but that can otherwise tolerate (short) interruptions.
Finally, at the top of the pyramid, where continuity of the process is truly non-negotiable, there’s the dynamic voltage restorer (DVR) which is a device that injects the missing voltage during a sag and keeps the load running, thus riding through the event. This stuff is so expensive that you only apply it if the cost of an unplanned shutdown is greater than the cost of the device divided by the time the plant is expected to operate.
Harmonics Are A Separate Problem
VFDs can be found in various heavy industries, and they are both sensitive to voltage fluctuations and a primary source of harmonic distortion that causes stress to connected equipment. IEEE 519 provides the relevant benchmarks as it establishes recommended limits for the total harmonic distortion (THD) measured at the point of common coupling.
Measuring the actual THD level against these limits is important as harmonic current behaves differently from a voltage sag. It rather manifests as transformer overheating, nuisance neutral-ground faults, or capacitor banks failing to improve power factor and instead forming a resonant circuit. Power factor correction equipment, if not properly dimensioned taking into account the harmonic environment, can worsen the situation. The solutions available – line or load reactors, passive filters, and active harmonic filters – depend on the source of harmonic current, which leads back to the initial recommendation to measure before taking any actions.
Maintenance Is Mitigation, Not An Afterthought
Much of what may appear as a random power quality issue is in truth a maintenance problem in a different guise. For instance, a loose high-voltage connection creates heat and resistance long before it fails but, from a power quality perspective, it’s a brief, isolated event. Similarly, termination problems occur slowly over time often due to something as simple as poor workmanship over tightening or under tightening connections.
Detection methods including thermal imaging, insulation resistance testing, partial-discharge, and evaluation of dielectric dissipation factor provide evidence of this form of breakdown long before it shows up on a PQ analyzer as a seemingly random transient. Loose connections in particular also highlight the need for solid maintenance of this equipment. As with stuck or inoperable cooling equipment, pay a little attention now with torque verification or contact/tension testing, and a lot less later on additional maintenance. This is especially true when you consider that a bad connection is both intermittent in nature and can lead to or be a result of loose material lying in a termination.
Here’s the thing: this is also where the work stops being something a general plant fitter should be doing. HV equipment carries risks that go well beyond a standard low-voltage isolation – arc flash energy, switching sequences, and earthing procedures that have to be followed exactly. Commissioning, testing, and ongoing inspection of this equipment is specialist work, and most facilities bring in certified high voltage electricians rather than relying on internal maintenance staff who aren’t licensed for HV isolation and switching. That’s not a compliance formality – it’s the difference between a fault being found on a scheduled inspection and it being found the hard way, mid-shift.
Design Decisions That Reduce Exposure Before Day One
Some of the most effective mitigation occurs even before you think about putting in any surge protective devices (SPDs) at all. Redundant feeder and dual-bus topologies mean a fault on one supply path doesn’t take the whole process line down with it – the load transfers to the alternate feeder and production keeps running while the fault gets isolated and cleared. Segregating sensitive loads from heavy motor-starting circuits, meanwhile, reduces the chance that a large motor start on one bus drags down voltage on a bus feeding PLCs and instrumentation elsewhere in the plant.
Proper earthing again shows up here as foundational, not optional. The connection of all your plant and equipment back to earth is the most boring, cheap sheet metal item in the bill of materials – and it’s the foundation for every single surge protection and power conditioning decision you’ll have to make downstream from it. Plants that get this wrong at the design stage end up trying to solve a topology problem with equipment, which is slower and more expensive than solving it with layout.
Safety and The Human Factor
None of this happens without strict process. Reliability-centered maintenance (RCM) frameworks prioritize which HV assets get inspected and how often based on failure consequence, not just a fixed calendar – a feeder supplying a single non-critical load doesn’t get the same attention as the bus feeding the main process line.
Every piece of that work runs under permit-to-work systems, documented switching procedures, and arc flash boundaries that dictate exactly what PPE is required and how close anyone can get to live equipment. Standards like AS/NZS 3000 and AS/NZS 4836 exist precisely because HV work has consequences that low-voltage electrical work doesn’t, and they’re not treated as guidelines – they’re the baseline for every job.
Closing The Loop With Data
The plants that get ahead of this problem log every power event against their process and downtime records, not just their electrical logs. Correlating a sag at 2:14pm with a scrapped batch or a line stoppage turns a vague sense that “the power’s been flaky” into a specific business case for a specific piece of equipment.
That trended data also feeds back into the maintenance schedule itself. If PQ logs show recurring stress on a particular feeder or MCC, that asset moves up the priority list in the CMMS, and inspection frequency adjusts accordingly. Industrial maintenance built this way stops being reactive – it becomes a feedback loop where the electrical data, the production data, and the maintenance plan all point at the same set of decisions.
Power fluctuations aren’t going away, and no facility eliminates them entirely. What separates plants that manage this well from plants that keep getting surprised is the sequence: measure first, protect in layers, maintain what you’ve installed, and put the work in the hands of people licensed to do it properly. Skip any one of those steps and the rest of the stack stops paying for itself.
