Power Factor Correction Calculator
Power factor correction capacitor sizing is a standard industrial electrical engineering calculation with rapid payback -- typically the fastest-return energy investment a facility...
Formula
Source: IEEE 18 (Shunt Power Capacitors), IEEE 1036 (Application of Shunt Capacitors), NEC 460 | Last reviewed: July 26, 2026
Examples
500 kW
= 277 kvar
- pf_old = 0.75
- pf_new = 0.95
500 kW: PF 0.75 to 0.95 needs 277 kVAr capacitor bank
100 kW
= 37.8 kvar
- pf_old = 0.82
- pf_new = 0.95
100 kW: PF 0.82 to 0.95 needs 38 kVAr - select 50 kVAr standard
1000 kW
= 595 kvar
- pf_old = 0.7
- pf_new = 0.92
1000 kW: PF 0.70 to 0.92 needs 595 kVAr - select 600 kVAr
Quick Reference Table
| Existing PF | Target PF 0.85 | Target PF 0.90 | Target PF 0.92 | Target PF 0.95 | Target PF 0.98 | Target PF 1.00 |
|---|---|---|---|---|---|---|
| 0.60 | 0.713 | 0.849 | 0.907 | 1.005 | 1.192 | 1.333 |
| 0.65 | 0.549 | 0.685 | 0.743 | 0.840 | 1.020 | 1.169 |
| 0.70 | 0.400 | 0.536 | 0.594 | 0.692 | 0.878 | 1.020 |
| 0.72 | 0.344 | 0.480 | 0.538 | 0.635 | 0.822 | 0.964 |
| 0.75 | 0.262 | 0.398 | 0.456 | 0.553 | 0.740 | 0.882 |
| 0.78 | 0.182 | 0.318 | 0.376 | 0.473 | 0.660 | 0.802 |
| 0.80 | 0.130 | 0.266 | 0.324 | 0.421 | 0.608 | 0.750 |
| 0.82 | 0.078 | 0.214 | 0.272 | 0.369 | 0.556 | 0.698 |
| 0.85 | 0.000 | 0.135 | 0.194 | 0.291 | 0.477 | 0.620 |
| 0.88 | -- | 0.059 | 0.117 | 0.214 | 0.401 | 0.543 |
| 0.90 | -- | 0.000 | 0.058 | 0.155 | 0.342 | 0.484 |
| 0.92 | -- | -- | 0.000 | 0.098 | 0.284 | 0.426 |
| 0.95 | -- | -- | -- | 0.000 | 0.186 | 0.328 |
| Total kVAr | Fixed/Auto | Steps | Typical Application | Approx Cost (USD) |
|---|---|---|---|---|
| 50 | Fixed | 1 | Individual motor < 100 HP | $1,200-1,800 |
| 100 | Fixed | 1 | Individual motor 100-200 HP | $2,000-3,000 |
| 150 | Auto | 3 x 50 | Small MCC bus correction | $5,000-7,000 |
| 300 | Auto | 6 x 50 | Medium industrial panel | $9,000-13,000 |
| 600 | Auto | 6 x 100 | Large industrial service | $17,000-24,000 |
| 1200 | Auto | 6 x 200 | Heavy industrial main swgr | $30,000-42,000 |
| 2000 | Auto | 8 x 250 | Very large facility > 3 MW | $48,000-65,000 |
Where is this used?
(1) Utility penalty elimination -- most utilities charge a penalty for PF below 0.85-0.90, either as a kVA demand charge or an explicit PF penalty multiplier.
Installing capacitors to reach PF = 0.95 eliminates this penalty entirely.
The annual savings are directly visible on the next month's bill and continue for the 20+ year life of the capacitors.
A mid-sized industrial plant with 3,500 kW average demand at 0.73 PF paying $9.50/kVA-month sees demand charges drop from (3,500/0.73) x $9.50 x 12 = $546,575 to (3,500/0.95) x $9.50 x 12 = $420,000 -- a $126,575 annual savings on a $65,000 capacitor installation (payback = 6.2 months).
(2) Transformer and switchgear capacity release -- improving PF from 0.70 to 0.95 reduces kVA demand by 26.3%.
For a facility with a 2,500 kVA transformer loaded at 85% (2,125 kVA) at PF = 0.70, the kW served is 1,488 kW.
After correction to PF = 0.95, the same transformer can serve 2,375 kW -- a 60% increase in real power capacity.
This often eliminates the need for a transformer upgrade when adding production equipment, saving $50,000-150,000 in capital cost.
(3) Reduced I squared R losses in facility distribution -- the reactive current component causes real heating in cables, bus bars, disconnects, and transformer windings.
At PF = 0.70, the total current is 1.43x the real-power current.
At PF = 0.95, it drops to 1.05x.
Since losses scale with I squared, the loss ratio is (1.43/1.05) squared = 1.86x -- losses nearly double at low PF.
For a facility with $800,000 annual electricity cost, distribution losses typically represent 2-3% = $16,000-24,000/year.
PF correction can eliminate 30-45% of these losses = $5,000-10,000/year additional savings.
(4) Voltage improvement -- reducing reactive current flow through the facility's distribution impedance reduces the voltage drop from the service entrance to the load.
The approximate voltage rise from capacitor installation: delta-V% = (kVAr_cap x X_trafo%) / kVA_trafo.
For a 500 kVAr capacitor on a 2,500 kVA transformer with 5.75% impedance, the voltage rise is (500 x 5.75) / 2500 = 1.15%.
While modest, this improvement at the motor terminals increases motor starting torque (proportional to V squared) and reduces the likelihood of nuisance contactor dropout during large motor starts.
(5) Generator capacity enhancement -- a standby generator rated 800 kW at 0.8 PF (1,000 kVA) can deliver 800 kW only if the load PF is at least 0.8.
If a facility's load is 700 kW at 0.65 PF = 1,077 kVA, it exceeds the generator's kVA rating even though the kW is below the engine limit.
Adding 450 kVAr of capacitors at the load bus reduces the kVA to 737 kVA -- within the generator rating -- and the full 700 kW can be served.
Without PF correction, the facility would need a 1,350 kVA generator (approximately 1,080 kW at 0.8 PF) -- a significantly more expensive machine.
(6) Renewable energy integration -- solar PV inverters can be configured to supply or absorb reactive power (Volt-VAR control per IEEE 1547-2018).
A facility with PV and a PF problem can use the inverter's reactive power capability as a dynamic VAR source, supplementing or replacing fixed capacitors.
However, inverters' reactive power capability reduces their real power output -- a 500 kW inverter providing 200 kVAr of reactive support can only deliver sqrt(500 squared - 200 squared) = 458 kW of real power.
The economics of using inverter kVAr vs capacitor kVAr depend on the value of the displaced solar kWh.
(7) Motor replacement -- when old motors are replaced with premium-efficiency units, the new motor often has a better PF (e.g., 0.85 vs 0.78 for a 100 HP motor).
The improved PF may reduce the required capacitor kVAr at the service, or in combination with capacitor correction, may achieve target PF with fewer capacitor steps.
Motor PF data should be factored into PF correction studies when planning equipment replacement.
(8) Green building certification -- LEED and other green building standards award points for energy performance optimization.
PF correction, while not directly reducing kWh consumption (except loss reduction), reduces the peak kVA demand on the grid and contributes to electrical infrastructure efficiency -- a metric increasingly tracked in sustainability reporting.
Real-World Usage Scenarios
Automated capacitor bank for a variable-load automotive assembly plant
An automotive stamping plant has a highly variable load profile: 2,400 kW during full production (three shifts), dropping to 300 kW during weekend maintenance. Measured PF varies from 0.68 at full load (large hydraulic presses) to 0.45 at light weekend load (mostly idling conveyors and lighting). A single fixed capacitor bank sized for full load would severely over-correct during light load, creating leading PF that could cause overvoltage on the 480V distribution and damage VFD front-end diode bridges. The solution: a 2,000 kVAr automatic bank with 8 steps of 250 kVAr, with a PF controller set to target 0.96 with a minimum lagging bias (never leading). During full production, 6-7 steps are engaged. During weekend light load, only 1-2 steps are needed. The controller's C/K setting (sensitivity) prevents hunting (excessive step switching). Installed cost: $72,000. Annual demand savings at $14/kVA-month: (2,400/0.68 - 2,400/0.96) x $14 x 11 productive months = $141,176. Payback: 6.1 months.
Detuned capacitor bank avoids resonance disaster at a plastic extrusion facility
A plastic extrusion plant with 45 VFDs (totaling 3,800 HP) and a PF of 0.71 on a 4,800 kW load engages an engineering firm to specify PF correction. The straightforward calculation indicates 3,100 kVAr needed (0.71 to 0.96). However, a harmonic survey reveals: THDv = 6.2% at the 480V bus (exceeds IEEE 519 limit of 5%), with strong 5th harmonic (4.8%) and 7th harmonic (3.1%). The short-circuit MVA at the bus is 55 MVA. Without detuning, a 3.1 MVAr capacitor bank would resonate at f_r = 60 x sqrt(55/3.1) = 60 x 4.21 = 253 Hz -- dangerously close to the 5th harmonic (300 Hz). Estimated amplification factor Q = 12, meaning 4.8% 5th harmonic voltage would produce 58% harmonic current in the capacitor -- far exceeding the capacitor's 135% continuous overcurrent rating (IEEE 18). The firm specifies a detuned bank with 7% series reactors, shifting resonance to 227 Hz (below 5th). The reactors add $31,000 to the project cost ($148,000 total vs $117,000 undetuned), but prevent a near-certain catastrophic failure. Additionally, the detuned bank absorbs some 5th harmonic current (acting as a passive filter), reducing bus THDv from 6.2% to 3.8% -- bringing the facility into IEEE 519 compliance.
Point-of-use capacitors at large motors reduce feeder losses in a cement plant
A cement plant has three 600 HP 4160V raw mill motors (NEC FLC = 69 A each, PF = 0.82 measured at terminals) located 1,200 ft from the main 4160V switchgear. The feeder cables (3 x #4/0 AWG, 15 kV EPR) for each motor carry 69 A of load current plus approximately 41 A of reactive current (PF = 0.82 means 57% real current, 43% reactive at a phase angle of 34.9 degrees). I squared R losses in each 1,200 ft feeder (R = 0.063 ohms/1000 ft x 1.2 = 0.0756 ohms per phase) = 3 x 69 squared x 0.0756 = 1,080 W per motor. Across three motors operating 8,000 hr/yr at $0.08/kWh, annual feeder loss cost = 1,080 x 3 x 8,000 / 1000 x $0.08 = $2,074. The plant engineer specifies 100 kVAr capacitors at each motor terminal (controlled by the motor starter auxiliary contact). This cancels the motor's reactive current at the motor, reducing feeder current from 69 A to 69 x 0.82 = 56.6 A (real current only). New feeder losses = 3 x 56.6 squared x 0.0756 = 727 W per motor. Savings = (1080-727) x 3 x 8000 / 1000 x 0.08 = $678/yr in feeder losses, plus 300 kVAr of main bus PF improvement (reducing demand charges). Capacitor cost: 3 x $2,800 (100 kVAr, 5 kV rated, including fused disconnect) = $8,400. Payback from loss savings alone = 12.4 years -- marginal. But combined with the demand charge reduction (300 kVAr at $12/kVA-month = $43,200/yr), total savings = $43,878/yr, payback = 2.3 months.
Common Mistakes to Avoid
Selecting capacitor kVAr based solely on the utility billing PF instead of a measured load profile
The utility bill reports PF as the average over the billing period (typically monthly), based on the accumulated kWh and kVARh registers. This average masks significant variation: a facility may have a monthly average PF of 0.85, but during second-shift light production the PF drops to 0.55 because large motors idle. A capacitor bank sized for the 0.85 average will be severely undersized during the low-PF periods and over-sized when PF is near unity. The correct approach: install a power quality analyzer at the main service entrance for a minimum of 7 days (covering one full production cycle, including weekend if applicable) to capture the PF variation pattern. Size the capacitor bank for the average kW and the minimum (worst) PF during normal operating hours, not the billing period average. The PF controller will manage the variation -- but the bank must have enough total kVAr capacity for the worst-case condition.
Installing capacitors without a harmonic survey in a VFD-dense facility
Capacitors and system inductance form a parallel resonant circuit. In facilities where VFDs, UPS systems, DC drives, and LED lighting drivers represent more than approximately 15-25% of total connected kVA, the harmonic currents at the 5th (300 Hz), 7th (420 Hz), 11th (660 Hz), and 13th (780 Hz) harmonics are significant. A capacitor sized for PF correction may create a resonance exactly at one of these frequencies. The result: harmonic currents are amplified (not damped) by the capacitor, causing: (a) capacitor failure from dielectric overheating (rated for 135% of rated current per IEEE 18 -- harmonics can push current to 200-300%), (b) transformer overheating from harmonic flux, (c) nuisance tripping of circuit breakers and relays from distorted voltage zero-crossings, and (d) in the worst case, resonant overvoltage causing insulation breakdown in cables and motor windings. Always perform a harmonic survey (minimum 7 days, capturing all operating modes) before specifying capacitors. If THDv > 3% or TDD > 15%, specify detuned banks with series reactors.
Neglecting to discharge capacitors before maintenance -- leading to arc flash injuries
Capacitors store energy as DC voltage after disconnection. A 300 kVAr 480V capacitor bank charges to the peak of the AC waveform -- approximately 480 x sqrt(2) = 679V DC. NEC 460.6 requires internal discharge resistors that reduce residual voltage to less than 50V within 1 minute. However: (a) the discharge resistors can fail open (this is the most common capacitor failure mode, with no external indication), (b) the 1-minute discharge is sufficient for code compliance but not for immediate maintenance access -- 50V can still produce a hazardous arc if shorted, and (c) a capacitor discharged to 0V can 'recover' voltage due to dielectric absorption (a phenomenon where the dielectric material releases stored charge minutes or hours after discharge). The safe practice: after de-energizing a capacitor bank, wait 5 minutes for the internal resistors to discharge, then verify zero voltage with a properly rated voltmeter (Cat III or Cat IV). Apply a temporary shorting jumper across the capacitor terminals before touching them. Document this procedure in the facility's electrical safety program per NFPA 70E.
Industry Standards Referenced
Frequently Asked Questions
How do I calculate the kVAr of a capacitor bank?
kVAr = kW x [tan(arccos(PF_old)) - tan(arccos(PF_new))]. Gather: (1) average or peak kW demand from utility bills or a 7-day power analyzer log (use average, not peak, to avoid over-sizing), (2) existing PF from the utility meter or power analyzer (use the minimum PF during normal operations, not the billing period average), (3) target PF (typically 0.95 -- this eliminates penalties in all utility jurisdictions and provides margin against PF degradation). Use this calculator for the exact value. Always round up to the next standard capacitor size: 25, 50, 75, 100, 150, 200, 250, 300, 400, 500 kVAr etc. For automatic banks, specify 4-8 equal steps for smooth regulation -- a 600 kVAr bank with 6 x 100 kVAr steps provides good resolution. If the facility has VFD loads, UPS, or other non-linear loads > 25% of total, a harmonic survey is mandatory before specifying the capacitor bank. Detuned banks with series reactors (typically 7% impedance) may be required.
Do capacitor banks save energy?
Yes, in two ways: (1) Reduced kVA demand charges -- most utilities charge for peak kVA demand ($/kVA-month), not just kWh energy. Improving PF reduces kVA directly, and this savings appears on the NEXT bill after installation and continues for the 20+ year life of the capacitors. (2) Reduced I squared R losses -- the facility's own conductors, transformers, and switchgear carry less reactive current, reducing heating losses. At PF = 0.70, the total current is 1.43x the real-power current; at PF = 0.95, it drops to 1.05x. Since losses scale with I squared, the loss reduction is approximately (1.43/1.05) squared = 1.86x. For a facility spending $500,000/year on electricity, distribution losses of 3% = $15,000/year. A 40% loss reduction saves $6,000/year -- secondary to demand charge savings but not negligible. The combined savings typically pay for the capacitor installation in 6-18 months. At the utility level, PF correction reduces transmission losses and frees up grid capacity -- that is why utilities incentivize or penalize it. Note: capacitors do not reduce the kWh drawn by the load itself -- a 100 kW motor still draws 100 kW of real power regardless of PF.
What happens if I over-correct (PF > 1.0)?
Leading power factor (capacitive, where the capacitor kVAr exceeds the inductive kVAr of the load) can cause voltage rise on the distribution system -- the Ferranti effect, where capacitive current flowing through inductive system impedance produces a voltage rise that can exceed equipment voltage ratings. On a lightly loaded 480V bus, over-correction of 15-20% can produce a 5-8% voltage rise, potentially damaging VFD DC bus capacitors (rated typically for 760-800V DC, corresponding to approximately 540-565V AC line -- a 480V system at +8% is 518V AC, well within rating but approaching limits for older equipment). Most utilities penalize leading PF (below -0.95 or equivalent) as well as lagging PF -- a leading PF condition means the customer is exporting reactive power to the grid, which causes voltage regulation issues for the utility. Automatic PF controllers prevent over-correction by disconnecting capacitor steps when reactive power goes capacitive. For fixed banks, ensure the capacitor kVAr does not exceed the minimum inductive kVAr draw of the facility at the lightest load condition -- typically the no-load magnetizing kVAR of the service transformer(s) plus the minimum connected motor magnetizing current. NEC 460.8 limits motor-connected capacitors to a kVAr rating that does not raise the motor's no-load PF above unity.
Should I install capacitors at the service entrance or at individual motors?
The optimal location depends on the facility layout and the source of low PF. Service entrance (main switchgear) -- simplest installation, single point of control, directly reduces utility measured kVA for billing purposes, lowest installed cost per kVAr. However, does not reduce I squared R losses in the facility's own feeders between the service entrance and the motors -- those feeders still carry the full reactive current. Best when: the low PF is distributed across many small motors (no single dominant load) and feeder runs from service to loads are short. Motor control center (MCC) -- corrects PF at the MCC bus, reducing reactive current in the main feeder from the service to the MCC. Moderate cost (one bank per MCC). Best when: a facility has multiple MCCs in different buildings or areas, with significant distance from the service. Individual large motor (>200 HP) -- capacitor connected directly at the motor terminals, switched with the motor starter. Minimum installed cost (no separate controller or enclosure needed) and maximum loss reduction (reactive current is cancelled at the source). NEC 460.8 compliance: capacitor kVAr must not exceed the value that would make the no-load PF unity (to prevent self-excitation overvoltage if the motor coasts down while capacitors remain connected). Best when: a small number of large motors (>200 HP) dominate the facility's PF, especially if motor feeders are long (>500 ft). Mixed strategy: individual capacitors at the 2-3 largest motors (cancelling 40-50% of the total kVAR) plus an automatic bank at the service for the remaining distributed load -- most cost-effective for large facilities.
How does harmonic distortion affect capacitor bank design?
Capacitor impedance decreases with frequency (X_C = 1 / (2 x pi x f x C)), while inductive reactance of transformers and cables increases with frequency (X_L = 2 x pi x f x L). At some harmonic frequency, these impedances become equal -- parallel resonance. At resonance, the circuit impedance becomes very large, and even small harmonic currents from VFDs or UPS produce large harmonic voltages. The resonance frequency is f_r = f_fundamental x sqrt(MVA_SC / MVAr_cap). If f_r is near the 5th harmonic (300 Hz), the 5th harmonic current (always present with 6-pulse VFD rectifiers) is amplified 5-15x -- overloading the capacitors (rated for 135% of rated current) and potentially causing rapid failure. Solution: detuned capacitor banks with series reactors (typically 7% impedance at fundamental, tuned to 189 Hz). These make the combined capacitor-reactor appear inductive at harmonics above the tuning frequency, eliminating resonance risk. The reactors add approximately 25-35% to the capacitor bank cost but are essential in any facility where VFDs, UPS, DC drives, or LED drivers exceed approximately 15-25% of total load. For severe harmonic environments (THDv > 7%), active harmonic filters that dynamically inject counter-phase harmonic currents are an alternative to (or complement of) detuned banks. A harmonic survey per IEEE 519 is the prerequisite for this design decision -- NEVER install a capacitor bank in a facility with non-linear loads without one.
Reviewed for accuracy
Reviewed against IEEE 18-2012 (Shunt Power Capacitors), IEEE 1036-2010 (Guide for Application of Shunt Power Capacitors), NEC 2023 Article 460, and IEEE 519-2014 harmonic limits · Last reviewed: July 26, 2026
All calculations are for reference only. Always verify with manufacturer data and a qualified engineer for critical applications. Learn about our editorial process.