How to Right-Size Power Factor Correction

Power factor correction is one of the best-understood fixes in industrial power: install capacitors to supply reactive power locally, raise your power factor, and eliminate the penalty on your utility bill. But "install capacitors" hides a lot of nuance, and getting the size wrong — in either direction — wastes money or creates new problems. This is a practical guide to right-sizing power factor correction, and why it has to start with real data.

A quick refresher

Motors, transformers, and other inductive loads draw reactive power (kVAR) to sustain their magnetic fields. This reactive power does no useful work, but the utility still has to deliver it, so they penalize low power factor — the ratio of useful (real) power to total (apparent) power. Correction works by installing capacitors that supply the reactive power locally, so the utility doesn't have to. Do it right and your power factor rises above the penalty threshold and the charge disappears.

The whole game is supplying the right amount of reactive power. Too little and you keep paying; too much and you cause new problems.

Why sizing from a single reading fails

The most common mistake is sizing correction from a one-time power-factor reading — a snapshot taken during a walk-through or pulled from a single bill. This fails because power factor and reactive demand vary constantly with load. Your plant's reactive power at 2 p.m. on a busy Tuesday is very different from 2 a.m. on a Sunday. Size your capacitors for one moment and you'll be wrong at every other moment.

Correct sizing requires trended data over real operating cycles — ideally days to weeks that capture your plant's full range of operating conditions: full production, partial load, and idle. Only then can you size correction that works across how the facility actually runs, not just how it looked at one instant.

The overcorrection trap

Sizing too small is the obvious error — you don't fully clear the penalty. But sizing too large is the subtler and more dangerous one. Overcorrection pushes your power factor leading (excess capacitance), which causes real problems:

• Some utilities penalize leading power factor too, so you can overshoot the penalty and land right back in one from the other direction.
• Voltage rise. Excess capacitance can raise system voltage, stressing equipment.
• Resonance and harmonics. Oversized or poorly applied capacitors can interact with harmonics in the system to create resonance, amplifying distortion and damaging capacitors and equipment — especially in facilities with lots of variable-frequency drives.

This is why "just add a big capacitor bank" is bad advice. The goal is to land in the correct window — above the penalty threshold, but not so far that you cause leading-PF problems.

Fixed vs. automatic correction

Two broad approaches, and the choice depends on how variable your load is:

• Fixed capacitors supply a constant amount of reactive power. Simple and cheap, and ideal for steady, predictable loads — for example, correcting an individual large motor that runs continuously.
• Automatic (switched) capacitor banks use a controller to switch capacitor stages in and out as your reactive demand changes, keeping power factor in the target window across varying load. More sophisticated, and the right choice for plants whose load swings widely through the day.

Trended data is what tells you which you need — and how many stages an automatic bank should have.

The right-sizing process

1. Measure first. Trend power factor, real power, and reactive power (kVAR) per meter over enough time to capture your full operating range.
2. Determine the target. You want to correct up to just above your utility's threshold (commonly 0.95), with margin — not all the way to unity, and never into significant leading territory.
3. Calculate the correction needed across your load range, accounting for how reactive demand varies.
4. Choose fixed vs. automatic based on that variability, and consider correcting at the load, at the panel, or at the service entrance depending on where the reactive power originates.
5. Account for harmonics. In drive-heavy facilities, work with someone who'll check for resonance risk and specify detuned/harmonic-rated capacitors if needed.
6. Verify after installation. Confirm the power factor now sits in the target window across operating conditions and the penalty is gone.
7. Keep watching. Capacitors fail silently and load changes over time — ongoing monitoring catches a failed stage before you slide back into penalties.

Why monitoring is essential before and after

Right-sizing is a data problem at both ends. Before: you need trended reactive-power data across real operating cycles to size correctly and avoid the overcorrection trap. After: you need continuous monitoring to confirm the correction is working and to catch the silent failure of a capacitor stage — a common event that can quietly return a plant to penalties without anyone noticing until the bill arrives. Sizing from a snapshot and then never checking again is how facilities end up either under-corrected, over-corrected, or back in penalties within a year.

A worked example

A motor-heavy plant is paying a power-factor penalty and gets a quote for a large fixed capacitor bank sized off a single reading. Before installing, they trend reactive power for two weeks and discover their load swings widely between shifts. A single fixed bank sized for the busy period would push them into leading power factor — and a new penalty — every night. Instead they install an automatic staged bank sized from the trended range, which keeps power factor in the target window across all conditions. Post-install monitoring confirms the penalty is gone and later flags a failed capacitor stage before it can cost them. Sizing from real data saved them from an expensive overcorrection mistake.

Common questions

Can't I just correct all the way to 1.0? No — aim for just above the threshold with margin. Correcting to unity leaves no room for load variation and risks tipping into leading power factor.

Fixed or automatic — which do I need? Steady load: fixed is fine and cheaper. Variable load: automatic staged banks. Your trended data decides.

Why did my correction stop working? Usually a failed capacitor stage (they degrade silently) or a change in plant load. Continuous monitoring catches both.

The bottom line

Power factor correction eliminates penalties — but only when it's right-sized, and right-sizing requires real trended data, not a single reading. Measure across your full operating range, correct to just above the threshold without overshooting into leading power factor, choose fixed or automatic based on your load's variability, and keep monitoring so a silent capacitor failure doesn't quietly undo your work.

OptimizeOS trends power factor and reactive power per meter — giving you the data to size correction right and the monitoring to keep it working.

Book a demo → or explore power quality →.