Quick answer: Power factor correction reduces the reactive power (kVAR) an inductive load draws by adding capacitors, which supply that reactive power locally instead of pulling it from the utility. You size the capacitor bank from the load's real power (kW) and the change in power factor: kVAR = kW × (tan φ1 − tan φ2), where φ1 is the existing and φ2 the target angle. Correcting to about 0.95–0.98 avoids utility penalties, frees up system capacity, and cuts losses — without over-correcting into a leading power factor. Size the bank with the power factor correction calculator.
What power factor actually is
Every AC electrical load draws two kinds of power. Real power (kW) does the useful work — it turns the motor, produces the light, generates the heat. Reactive power (kVAR) does no useful work; it energizes the magnetic fields in motors, transformers and other inductive equipment, sloshing back and forth between the load and the source every cycle. The vector sum of the two is apparent power (kVA), which is what the wires, transformers and utility actually have to carry:
kVA² = kW² + kVAR²
Power factor is the ratio of the useful power to the total: PF = kW ÷ kVA. A power factor of 1.0 (unity) means all the current is doing useful work. A power factor of 0.75 means the system is carrying 33% more current than the work requires, just to shuttle reactive power around. That extra current costs money — in utility penalties, wasted capacity, and higher losses.
Why a poor power factor costs you
- Utility penalties. Many commercial and industrial tariffs charge a penalty when power factor falls below a threshold (often 0.90 or 0.95), because the utility must size its system for the apparent power (kVA), not just the real power it bills.
- Wasted capacity. A transformer or feeder rated in kVA is partly consumed by reactive power. Correcting the power factor frees up real capacity — sometimes deferring an expensive service upgrade.
- Higher losses and voltage drop. More current means more I²R heating in cables and transformers, and more voltage drop to distant loads.
The culprit is almost always inductive load — lightly loaded motors, transformers, welders, and older fluorescent/HID lighting all draw reactive power and drag the power factor down.
How capacitors fix it
Inductive loads draw lagging reactive power. Capacitors draw leading reactive power — the exact opposite. Install capacitors near the inductive load and they supply the reactive power the motors need locally, so it no longer has to travel all the way from the utility through the transformers and cables. The reactive current effectively circulates between the capacitor and the motor over a short path, and the utility only has to supply the real power. That's power factor correction in a sentence: add capacitors to cancel the inductive kVAR.
The sizing formula
You size the capacitor bank (in kVAR) from the load's real power and how far you want to move the power factor:
Capacitor kVAR = kW × (tan φ1 − tan φ2)
where:
- kW = the real power of the load,
- φ1 = the angle of the existing power factor (cos φ1 = current PF), and
- φ2 = the angle of the target power factor (cos φ2 = desired PF).
The term (tan φ1 − tan φ2) is the amount of reactive power per kW you need to remove — it's tabulated in every power-factor multiplier table, but the power factor correction calculator computes it directly from your existing and target PF.
Worked example: correcting a plant's power factor
A facility has a real load of 500 kW at an existing power factor of 0.75 (lagging), and the utility wants at least 0.95.
- Existing: cos φ1 = 0.75 → tan φ1 = 0.882
- Target: cos φ2 = 0.95 → tan φ2 = 0.329
- Required capacitors: 500 × (0.882 − 0.329) = 500 × 0.553 = ≈ 277 kVAR
So a capacitor bank of about 275–300 kVAR raises the plant from 0.75 to 0.95. Check the effect on apparent power: before, kVA = 500 ÷ 0.75 = 667 kVA; after, kVA = 500 ÷ 0.95 = 526 kVA — a 141 kVA reduction, which frees transformer and feeder capacity and eliminates the penalty. Size the bank with the power factor correction calculator.
How much to correct: aim for 0.95–0.98, not 1.0
It's tempting to correct all the way to unity, but that's usually the wrong target. Correcting to about 0.95–0.98 captures nearly all the benefit (penalties disappear, capacity is freed) while leaving a safety margin. Pushing to 1.0 — or beyond, into a leading power factor — is risky: at light load, fixed capacitors can over-correct, driving the power factor leading, which raises voltage and can cause instability, resonance, and its own utility penalties (some tariffs penalize leading PF too). The last few points toward unity also cost disproportionately more capacitor kVAR for little gain. The sweet spot is a comfortable 0.95–0.98.
Fixed vs. automatic (switched) capacitor banks
| Type | How it works | Best for |
|---|---|---|
| Fixed capacitors | A set kVAR always connected, often at a specific motor | Steady loads; correcting individual large motors at the load |
| Automatic bank | A controller switches capacitor steps in/out to hold a target PF | Varying plant load — avoids over-correction at light load |
Where the load swings widely, an automatic bank is important: it adds capacitance as load rises and removes it as load falls, so you never over-correct into a leading power factor during light-load periods. Fixed capacitors are ideal at a single large motor (correcting it at the source), but a plant with variable load needs the automatic control.
The harmonics warning: detuned reactors
One serious caution: capacitors and the system's inductance form a resonant circuit, and if that resonance lands near a harmonic frequency present in the system (from VFDs, rectifiers and other non-linear loads), the result can be harmonic resonance — amplified currents that overheat and destroy the capacitors and disturb the system. In any facility with significant non-linear load, power-factor capacitors should be detuned (fitted with series reactors that shift the resonant point away from the troublesome harmonics). Ignoring harmonics is a common and expensive mistake — see our guide to neutral current and harmonics for why non-linear loads change everything, and coordinate limits with IEEE 519.
Sizing correction from your utility bill
You don't always start from a clean load calculation — often the trigger is a power-factor penalty on the utility bill, and you can size the correction straight from billing data. Utility bills for commercial and industrial accounts frequently list both kWh (real energy) and kVARh (reactive energy), or the billed power factor directly. From those you can back out the average power factor and the kW, then apply the same formula — kVAR = kW × (tan φ1 − tan φ2) — using your billed (existing) power factor and the utility's threshold (target) power factor. Sizing from real billing data has an advantage: it reflects the actual operating power factor over the billing period, not a nameplate assumption, so the correction is tuned to how the facility really runs. Just remember that billed power factor is an average — a facility with wide load swings still needs an automatic bank so it doesn't over-correct at the low-load end.
Correcting individual motors at the source
For large motors that run steadily, the most effective correction is a fixed capacitor connected right at the motor terminals (or its starter), sized to the motor's magnetizing (no-load) reactive power. This is elegant because the capacitor switches with the motor — no separate controller needed — and it relieves reactive current all the way back through the motor's feeder. There is one firm rule: size the capacitor to the motor's no-load kVAR, not more. An oversized capacitor on a motor can cause self-excitation when the motor is switched off while still spinning — the capacitor drives the motor as a generator, producing dangerous over-voltages that damage the motor and capacitor. Motor manufacturers and NEMA publish maximum capacitor kVAR by motor size and speed for exactly this reason. For steady, large motors, terminal correction is efficient; for the plant as a whole with varying load, use a central automatic bank.
The capacity-release benefit
Beyond avoiding penalties, power factor correction frees real capacity in your electrical system, and this is often the bigger prize. Because your transformers and feeders are rated in kVA, reactive power consumes part of their capacity. Raising the power factor lowers the kVA for the same real load, so the same equipment can carry more useful load — sometimes deferring an expensive service or transformer upgrade. In the worked example, correcting 500 kW from 0.75 to 0.95 dropped the apparent power from 667 kVA to 526 kVA, a 141 kVA reduction. If that plant was near the limit of a 600 kVA transformer, the correction alone created room to add load without upsizing the transformer. Quantify this against the transformer sizing calculator — the freed kVA is real, usable capacity.
Economics and payback
Power factor correction is one of the more attractive electrical investments because it attacks a recurring cost. The payback comes from three streams: eliminated penalties (immediate and ongoing), reduced demand charges where the utility bills on kVA demand, and lower losses in the facility's own cables and transformers (less current = less I²R heating). For a facility paying monthly power-factor penalties, a capacitor bank often pays for itself in months to a couple of years, after which the savings are pure return. The economics improve further if the freed capacity defers a capital upgrade. The main cost risks — and reasons to involve an engineer — are getting the harmonics right (detuning) and avoiding over-correction, both of which turn a good investment into an expensive problem if ignored.
Common power factor correction mistakes
- Over-correcting to leading PF. Aim for 0.95–0.98; leading power factor raises voltage and can incur its own penalties.
- Fixed capacitors on a variable load. They over-correct at light load — use an automatic switched bank.
- Ignoring harmonics. Undetuned capacitors in a VFD-heavy plant can resonate and fail; use detuned banks.
- Correcting kVAR from kVA instead of kW. The formula uses real power (kW) and the tangents of the angles.
- Placing correction only at the service. Capacitors at the load relieve the cables too; service-level correction only fixes the utility bill.
- Forgetting discharge and protection. Capacitor banks need discharge resistors and proper overcurrent protection.
Passive capacitors vs. active correction
Traditional power factor correction uses passive capacitor banks — simple, proven, and cost-effective for correcting the steady, displacement power factor caused by ordinary inductive loads. But in facilities dominated by non-linear electronics (large VFD populations, data centers, big rectifier loads), the “bad” power factor isn't purely the lagging displacement kind — a lot of it comes from harmonic distortion, and capacitors alone can't fix that (and can be endangered by it). Two more advanced tools address this:
- Detuned (reactor-connected) capacitor banks — still passive, but with series reactors that both shift the resonant point away from harmonics and provide modest harmonic filtering.
- Active harmonic filters / active VAR compensators (and STATCOMs) — power-electronic devices that inject corrective current in real time, correcting displacement power factor and cancelling harmonics dynamically as the load changes. They cost more but handle rapidly varying, harmonics-rich loads that would trouble a switched capacitor bank.
The right choice depends on the load. A plant of ordinary motors is a textbook case for passive capacitors (fixed at large motors, automatic bank for the varying remainder). A modern facility full of drives and electronics usually needs detuned banks at minimum, and sometimes active filtering — which is why the harmonic assessment (per IEEE 519) should come before you pick the correction hardware, not after. Matching the correction technology to the actual nature of the poor power factor is the difference between a durable installation and a bank that overheats and fails.
Standards and references
| Reference | What it covers |
|---|---|
| IEEE 519 | Harmonic limits — drives the need for detuned banks |
| NEC Article 460 | Capacitors — conductor sizing, overcurrent, discharge |
| IEEE 1036 | Application guide for shunt power capacitors |
| Formula | kVAR = kW × (tan φ1 − tan φ2); PF = kW ÷ kVA |
The bottom line
Power factor correction adds capacitors to supply reactive power locally, cutting the current the system carries, freeing capacity and eliminating utility penalties. Size the bank from kVAR = kW × (tan φ1 − tan φ2), correct to a comfortable 0.95–0.98 (never over-correct into a leading power factor), use an automatic bank where the load varies, and detune the capacitors in any harmonics-rich plant. Size it with the power factor correction calculator, check the freed capacity against the transformer sizing calculator, and have the installation designed by a licensed electrical engineer per NEC 460 and IEEE 519.
Frequently asked questions
How do you calculate power factor correction?
Size the capacitor bank from the load's real power and the change in power factor: kVAR = kW times (tan of the existing power-factor angle minus tan of the target angle). For example, 500 kW at 0.75 power factor corrected to 0.95 needs 500 times (0.882 minus 0.329), about 277 kVAR. The term in brackets is the reactive power removed per kW; a calculator computes it directly from your existing and target power factor.
How do capacitors improve power factor?
Inductive loads such as motors and transformers draw lagging reactive power to energize their magnetic fields. Capacitors draw leading reactive power, the exact opposite, so installing them near the load supplies that reactive power locally instead of pulling it all the way from the utility. The reactive current then circulates over a short path between the capacitor and the motor, the utility only supplies the real power, and the overall power factor rises toward unity.
What power factor should you correct to?
Aim for about 0.95 to 0.98, not 1.0. That captures nearly all the benefit - penalties disappear and capacity is freed - while leaving a safety margin. Correcting all the way to unity or beyond risks over-correction into a leading power factor at light load, which raises voltage and can cause instability, resonance and its own utility penalties. The last points toward unity also cost disproportionately more capacitor kVAR for little gain.
What is a leading power factor and why is it bad?
A leading power factor happens when there is more capacitance than the inductive load needs, so the net reactive power reverses direction. Fixed capacitors sized for full load can over-correct at light load and drive the power factor leading, which raises system voltage, can cause instability and resonance, and may incur its own utility penalties since some tariffs penalize leading power factor. Automatic switched banks avoid this by removing capacitance as load falls.
What is the difference between fixed and automatic capacitor banks?
A fixed capacitor bank keeps a set kVAR always connected, ideal for a steady load or for correcting a single large motor right at the source. An automatic bank uses a controller to switch capacitor steps in and out to hold a target power factor as the plant load varies, which prevents over-correction into a leading power factor during light-load periods. Facilities with widely varying load need the automatic type.
Do harmonics affect power factor capacitors?
Yes, significantly. Capacitors and the system inductance form a resonant circuit, and if that resonance lands near a harmonic frequency produced by VFDs, rectifiers and other non-linear loads, harmonic resonance can amplify currents and destroy the capacitors. In any plant with significant non-linear load, power-factor capacitors should be detuned with series reactors that shift the resonant point away from the harmonics. Ignoring harmonics is a common and costly mistake.
Where should power factor correction capacitors be installed?
They can be installed at the service (fixing the utility bill only), at a distribution panel, or right at individual large motors. Correcting at or near the load is the most effective because it relieves the reactive current from the cables and transformers feeding that load, not just the utility connection. Service-level correction eliminates the penalty but does not free up capacity in the downstream distribution the way load-level correction does.
Is a power factor correction calculator accurate for design?
A calculator that applies kVAR = kW times (tan1 minus tan2) gives the correct capacitor bank size to move from an existing to a target power factor. A complete design also requires deciding between fixed and automatic banks, detuning for harmonics per IEEE 519, capacitor conductor sizing, overcurrent protection and discharge means per NEC Article 460, and checking for over-correction at light load, all verified by a licensed electrical engineer.