How is a power factor penalty calculated?

Most utilities bill a penalty in one of three ways: a charge applied to the ratio of kVA to kW demand, a direct charge per kVAR of reactive demand above a threshold, or an adjustment to the demand charge based on how far power factor falls below a target such as 0.95. The common structure is a ratchet — no penalty above 0.95, a moderate surcharge between 0.90 and 0.95, and a steeper rate below 0.90.

Is 0.95 power factor good?

0.95 is the threshold most utilities use as the point where penalties begin, so it is the minimum rather than a good target. Operating at 0.98 to 0.99 captures nearly all available savings and leaves margin for load changes. Correcting beyond 0.99 risks leading power factor at light load, which some tariffs also penalise.

Can I fix power factor without a capacitor bank?

Yes. Active front ends on drives, synchronous condensers, and static var generators all provide reactive compensation without capacitors, and none of them risk harmonic resonance. They cost more per kVAR than capacitors but are the better answer where harmonic distortion is already present, because a capacitor bank added to a distorted network can amplify harmonics rather than improve power factor.

Why did my power factor drop after installing VFDs?

A drive with a diode rectifier front end draws current in pulses, so its displacement power factor is high but its true power factor is low. Utilities that meter true power factor will bill for that difference. The distortion component is not correctable with capacitors, which is why VFD-heavy plants often see penalties persist despite a correctly sized capacitor bank.

Billing

How power factor penalty is calculated

The charge is not arbitrary. It follows from how the utility measures your demand — and once you know which structure applies, the payback period becomes arithmetic.

A power factor penalty is a billing adjustment applied when the ratio of real power (kW) to apparent power (kVA) falls below a threshold set in the tariff, commonly 0.95. The utility must supply and carry the extra current needed for reactive power, so it recovers that cost from the customer causing it.

Three structures account for most tariffs: a multiplier on the demand charge, a direct charge per kVAR of excess reactive demand, or a ratchet that steps the surcharge as power factor drops.

Updated 22 October 20268 min readFigures in USD, illustrative rates

Why the utility charges for it

Real power does work. Reactive power establishes the magnetic field that motors, transformers and ballasts need in order to operate, and it is not consumed — it sloshes back and forth between the load and the source sixty times a second, or fifty.

The problem is that the utility must size cables, transformers and generators for the total current, and that current includes the reactive component. A plant drawing 500 kW at 0.75 power factor pulls 667 kVA through equipment that only needs to deliver 500 kW of useful work. The utility funds the difference and bills it back.

# the three quantities
kVA = √( kW² + kVAR² )
PF = kW / kVA
# at 500 kW and 0.75 PF
kVA = 500 / 0.75 = 667 kVA ← billed demand
kVAR = √(667² − 500²) = 441 kVAR

The three penalty structures you are likely to face

Structures vary by utility and country. Check the tariff sheet for the exact wording — the distinction between measured and estimated reactive demand matters.
Structure	How it is applied	Typical magnitude
Demand charge multiplier	Billed demand is calculated as kW ÷ PF, or kW × a multiplier that grows as PF falls.	+5% to +30% of demand charge
Reactive demand charge	A separate rate per kVAR, applied only to the portion above a threshold such as 50% of kW.	$0.50–$2.00 per kVAR
Stepped ratchet	No penalty above 0.95; increasing surcharge at 0.90–0.95 and steeper below 0.90.	0% / +2% / +5% or similar
Incentive (the inverse)	A credit applied for PF above 0.95, sometimes up to 0.99.	−1% to −3% of demand

The fourth row is worth checking before assuming the only outcome is a penalty. Some tariffs pay a credit for power factor above 0.95, which changes the economics of over-correction in the opposite direction.

A worked example

Take a plant with 500 kW maximum demand, operating at 0.75 power factor, on a tariff that charges $12 per kVA of billed demand and applies no separate reactive charge.

	Present (0.75)	After correction (0.95)	Difference
Maximum demand	500 kW	500 kW	—
Power factor	0.75	0.95	+0.20
Billed demand	667 kVA	526 kVA	−141 kVA
Monthly demand charge	$8,004	$6,312	−$1,692
Annual saving			$20,304
Compensation required		277 kVAR

At roughly $25–45 per kVAR installed for a conventional capacitor bank, the 277 kVAR installation costs somewhere between $7,000 and $12,000 and pays back in under eight months. That ratio is why power factor correction is usually the first project approved in an energy audit.

If the plant also has a reactive demand charge of $1.50 per kVAR above 50% of kW, the saving roughly doubles.

Sizing the compensation

The capacitor rating is not the difference between present and target kVAR measured directly — it is the difference between the tangents of the two power factor angles, multiplied by real power.

Q_c = P × ( tan φ₁ − tan φ₂ )
# P = real power (kW)
# φ1 = acos(present PF), φ2 = acos(target PF)

# 500 kW, 0.75 → 0.95
tan(acos 0.75) = 0.8819
tan(acos 0.95) = 0.3287
Q_c = 500 × (0.8819 − 0.3287) = 277 kVAR

Two corrections are almost always needed on top of the calculation. First, size against the maximum demand that will actually occur, not the annual average, because the penalty is applied at peak. Second, if harmonic distortion is present, add a detuning reactor and recalculate at the detuned voltage — a 7% reactor reduces the effective capacitor output by roughly 7%.

Use the kVAR calculator to run the numbers for your own figures.

Four traps that make the correction fail

Correcting to 0.99 or above. At light load the plant can go leading, which some tariffs penalise and which can cause voltage rise. Target 0.95–0.98.
Ignoring harmonics. A capacitor bank on a network with significant distortion can enter resonance and fail within months. Check the resonance order before specifying.
Sizing on annual averages. The penalty is triggered by the peak billing demand, so average load data systematically under-sizes the bank.
Step control that is too coarse. A single fixed bank cannot track a plant whose load swings between 20% and 100%. Automatic step control, or a static var generator, is needed in that case.

True versus displacement power factor

This distinction catches out plants that installed drives. Displacement power factor looks only at the phase shift between fundamental voltage and current. True power factor includes distortion, and a six-pulse rectifier can have a displacement factor near 0.98 while its true power factor sits at 0.70.

The practical consequence: if the utility meters true power factor, capacitors will not fix the shortfall, because capacitors cannot correct distortion. In that case the correct answer is filtration or an active front end, and a capacitor bank may make the situation worse.

Working out whether correction still pays?

Send us the last three utility invoices and the single-line diagram. We will identify which billing structure applies and model the payback before you commit to hardware.

Send the invoices