What is power factor?

The mechanics

Alternating-current loads draw two kinds of current. Real power, measured in kilowatts, does the work: it turns the shaft, heats the element, lights the lamp. Reactive power, measured in kilovolt-amperes reactive (kVAr), does not. It builds and collapses the magnetic fields that induction motors, transformers and ballasts need in order to function, and it flows back and forth between the load and the source without ever being consumed.

The two combine vectorially into apparent power in kilovolt-amperes:

kVA² = kW² + kVAr² Power factor = kW ÷ kVA

That is the whole of it. A facility with a power factor of 1.0 draws only real power. A facility at 0.80 is drawing 25% more current than its real load requires, and every conductor, transformer, breaker and utility feeder between the generator and the motor has to be sized for that larger current.

There are three power factors, and confusing them is the most common technical error in this subject:

• Displacement power factor — the phase angle between the voltage and current fundamentals. This is what induction motors cause and what capacitors correct.
• Distortion power factor — the effect of harmonic current, which is at a different frequency from the voltage and therefore contributes to apparent power while contributing nothing to real power. Rectifier front ends, variable-frequency drives, LED drivers and switch-mode supplies cause this.
• True power factor — the product of the two. This is what a meter actually reports and what a utility actually bills.

The distinction has money attached: a capacitor bank corrects displacement power factor and does nothing for distortion power factor. A plant that has converted most of its motors to VFDs may find its displacement power factor excellent and its true power factor still poor, and installing capacitors into that condition can create a resonant circuit with the drives' harmonic current and make things measurably worse.

How utilities charge for it

Three clause structures dominate. Which one applies is written into the tariff, not into the industry.

Clause type	How it works
Adjusted demand	Billed kW is scaled up when power factor falls below a threshold, typically 0.90 or 0.95. Billed kW = actual kW × (threshold ÷ actual PF).
kVA demand	The demand charge is levied on apparent power directly, so poor power factor raises the bill automatically with no separate penalty line.
Reactive charge	A stated rate per kVAr or per kVArh above an allowance, appearing as its own line item.

Power factor is assessed at the billed meter and usually on the coincident demand interval — the same interval that sets the demand charge — not as a monthly average. A facility can hold 0.94 all month and be penalised on the single interval where a large motor started unloaded.

Worked example

A plant is billed under an adjusted-demand clause with a 0.95 threshold at $14.50 per kW-month.

• Measured demand in the peak interval: 1,200 kW
• Measured power factor in that interval: 0.82
• Billed demand = 1,200 × (0.95 ÷ 0.82) = 1,390 kW
• Penalty portion = 190 kW × $14.50 = $2,755 per month
• Annualised: $33,060 per year

The apparent power in that interval was 1,200 ÷ 0.82 = 1,463 kVA, against 1,263 kVA if the plant had held 0.95. The 200 kVA difference is 137 kVAr of correction — roughly two 75 kVAr steps, at a cost that this arithmetic pays back inside a year at most sites.

Note what the example does not claim: correcting to 0.95 saves no kilowatt-hours at all. The energy charge is unchanged. The saving is entirely on the demand line.

What it takes to measure

Power factor is the highest-return power-quality outcome available and also the cheapest, because the binding constraint is low. It needs coincident kW and kVAr at the billed meter, at the utility's demand interval — usually 15 minutes — plus the specific power-factor clause from the tariff. Higher resolution adds nothing to the billing question.

It does need a real three-phase meter with its own voltage reference. A self-powered current sensor clamped to a single conductor has no voltage input and therefore cannot compute a phase angle; it cannot report true power factor at all, whatever the datasheet implies.

Current-transformer phase-angle error also matters far more here than it does for energy measurement, and it matters most exactly where power factor is worst. At a power factor of 0.45, a 3.3° CT phase error produces a 13.1% error in reported power. Split-core CTs typically run 1.0–2.0° of phase error, older units up to 6.0° — systematically worst on the lightly-loaded motors that dominate a poor power factor.

What power factor is not

It is not efficiency. A motor at 0.75 power factor is not 75% efficient. The two are unrelated quantities: efficiency is output work divided by input real power, and power factor is real power divided by apparent power. A high-efficiency motor running unloaded has excellent efficiency ratings and terrible power factor.

Correcting it does not reduce energy consumption. This is the most persistent misconception in the field, and it is usually introduced by whoever is selling the capacitors. Reactive current does cause a small amount of extra resistive loss in the facility's own conductors, but the recoverable kWh from that is a rounding error. The money in power factor correction is on the demand line and the penalty line, and a proposal that promises kWh savings from capacitors should be asked for its derivation.

A "power factor correction device" for a small building is usually not correcting anything. Residential and small-commercial customers are almost never billed on power factor, so there is no penalty to remove.

It is not a power quality problem in the IEEE sense. Low power factor is a billing and capacity condition. Voltage sags, harmonics and transients are power quality events. They travel together in practice and they are different things.