Power factor correction calculator

Use the calculator

Enter the load in kW, the existing power factor, the target power factor, voltage, phases, and capacitor connection. The calculator returns required kVAR, the per-phase µF capacitance for delta or wye, the next standard NEMA CP 1 step, and the line-current reduction percentage.

The four PF-correction formulas

The single design equation. Multiplied by P (the actual load), it scales the bank size to your installation. A 1 000 kW factory at PF 0.75 → 0.95 needs Q_correction = 1 000 × (0.882 − 0.329) = 553 kVAR. Round up to the next standard step (typically 600 kVAR in 4 × 150 kVAR or 6 × 100 kVAR switched stages).

Each capacitor in a delta-connected bank sees the full line-to-line voltage. For wye (star) connection, swap V_LL² for V_LN² where V_LN = V_LL/√3 — wye capacitors see ⅓ the voltage but need 3× the capacitance for the same Q. Most LV banks are delta; very large MV banks are double-wye for unbalance protection.

Correcting from PF 0.75 to 0.95 reduces line current by 1 − 0.75 / 0.95 = 21 % for the same real load. The conductor, transformer, and breaker all run cooler — sometimes enough to defer a service upgrade by a decade. The utility-side win is the kVA demand charge dropping in proportion.

How to size a capacitor bank, step by step

1. Measure or estimate the existing power factor. Read PF₁ from the utility bill or a clip-on power meter. If unavailable, use a typical value: pure motor loads run 0.80–0.90 at full load (lower when lightly loaded), datacenter / IT loads run 0.95+ thanks to active PFC, mixed industrial sites usually sit around 0.75–0.85.
2. Pick the target power factor. Most utilities require PF₂ ≥ 0.85 or 0.90 to avoid penalties. Optimal range is 0.92–0.95 — high enough to clear billing thresholds with margin, low enough to avoid over-correction (leading PF) at light load. Above 0.99 risks resonance with system harmonics and is rarely worthwhile.
3. Compute the required kVAR cancellation. Q_correction = P × (tan φ₁ − tan φ₂), where φ = arccos(PF). The calculator returns this in kVAR — the rating of capacitor bank you need to install. Round up to the next standard NEMA CP 1 step: 5, 7.5, 10, 15, 20, 25, 30, 40, 50, 60, 75, 100, 125, 150, 200, 250, 300 kVAR.
4. Convert kVAR to capacitance in µF. For 3-phase delta: C_per_phase = (Q/3) × 1000 / (2π·f·V_LL²). For 3-phase wye: same formula with V_LN = V_LL/√3. For single-phase: C = Q × 1000 / (2π·f·V²). The calculator handles all three configurations and returns both per-phase and total capacitance.
5. Decide on fixed vs switched capacitor banks. If the load is steady (datacenter, large continuous motor): use a single fixed bank. If the load varies (multiple motors cycling on / off, batch processing): use a 4–6 stage switched bank with a PF controller that switches steps as needed. Fixed is cheaper; switched avoids over-correction at low load.
6. Verify against harmonics and resonance. On harmonics-rich systems (VFDs, rectifiers, induction furnaces), an unprotected capacitor bank can resonate with the source impedance and amplify harmonic currents. Use detuned banks (capacitor + reactor in series tuned to ~190 Hz at 60 Hz / ~135 Hz at 50 Hz) to block resonance and provide partial harmonic filtering.

Reference values

Typical PF by load type

Use these as starting estimates when you don\'t have measured PF data.

Standard NEMA CP 1 capacitor steps

Round Q_correction UP to the next of these. Common bank topology: 4–6 equal steps controlled by a PF controller.

Worked example: 250 kW factory at PF 0.72 → 0.95

A medium-sized factory draws 250 kW at 0.72 lagging on a 400 V 3-phase 50 Hz system. The utility threatens a $4 / kVA demand-charge surcharge above PF 0.85. Compute the capacitor bank to bring PF to 0.95.

A 200 kVAR detuned switched bank costs roughly $8 000–$15 000 installed. Payback against the demand charge is 2–4 years — typical for industrial PF correction projects. Add the secondary benefit of cooler conductors and longer transformer life, and the real ROI is shorter.

Variants and special cases

Fixed vs switched banks

Fixed — single permanent capacitor bank. Cheap, simple. Risks over-correction at light load (leading PF, also penalised). Use only when load is steady. Switched — multi-stage bank (4–8 equal steps) with a PF controller that switches steps in/out as load varies. More expensive but maintains target PF across the load profile. Standard for any factory or commercial site with cycling loads.

Detuned vs undetuned

Plain capacitors resonate with system inductance and amplify harmonics from VFDs and rectifiers. Detuned banks add a small series reactor to each capacitor stage, tuning the bank below the lowest harmonic (typically to 134 Hz at 50 Hz / 189 Hz at 60 Hz). Detuned banks cost ~30 % more but are mandatory above ~10 % VFD load on the bus.

Active vs passive correction

Passive — capacitor banks (this calculator). Cheap, robust, but only corrects fundamental-frequency PF and can\'t handle harmonics. Active — IGBT-based static VAR generator (SVG) or active harmonic filter (AHF). Continuously variable kVAR, no resonance, also filters harmonics. 5–10× more expensive per kVAR but the only choice for highly non-linear loads or very fast-changing loads (e.g. arc furnaces).

Capacitor placement strategies

At motor terminals — best for large constant-running motors (≥30 kW); the cap cancels VAR right at the source and turns off with the motor. At the MCC bus — covers all motors on that MCC; one larger bank instead of many small ones. At the service entrance — corrects only the billable PF measured at the meter; doesn\'t unload internal feeders. Most large installations combine all three.

The IEEE 1036 sizing guidance

The reactive power required to correct the power factor from PF₁ to PF₂ at constant real power P is given by Q_c = P (tan φ₁ − tan φ₂), where φ = cos⁻¹(PF). The capacitor bank rating shall be selected as the next-larger standard kVAR step from this calculated value, with consideration for system harmonics, switching transients, and discharge resistor requirements per IEEE Std 18.

Related calculators and references

Frequently asked questions

What is power factor?

The cosine of the phase angle between AC voltage and current. PF = real power (kW) / apparent power (kVA). PF = 1 means current and voltage are perfectly in phase (purely resistive load). PF < 1 means current lags or leads voltage (inductive or capacitive load), so the alternator and conductor must carry more current than the actual work requires. Industrial loads typically run PF 0.7–0.85 lagging due to induction motors.

What is the formula for required kVAR?

Q_correction = P × (tan φ₁ − tan φ₂), where P is real power in kW, φ₁ is the existing phase angle (arccos of current PF), φ₂ is the target phase angle. Example: 100 kW load at PF 0.75 → tan φ₁ = 0.882; target PF 0.95 → tan φ₂ = 0.329. Q_correction = 100 × (0.882 − 0.329) = 55.3 kVAR.

How do I size a capacitor bank?

Compute Q_correction in kVAR (above), then convert to capacitance in µF: C (µF) = Q × 1000 / (2π·f·V²) × 10⁶. For 3-phase delta-connected banks, divide Q by 3 and use V_LL² for each capacitor. For wye-connected, use V_LN² = (V_LL/√3)². The calculator does this automatically. Round up to the next standard NEMA CP 1 step.

What does "leading" vs "lagging" power factor mean?

Lagging — current lags voltage, caused by inductive loads (motors, transformers). Most industrial sites are lagging. Leading — current leads voltage, caused by capacitive loads (over-correction with too much PF capacitor bank, long underground cables at light load). Both are penalised by some utilities. Target is unity (PF = 1) or slight lag (0.95).

Why does utility charge a PF penalty?

A 200 kW load at PF 0.7 draws the same line current as a 285 kVA load at PF 1.0 — the alternator, transformer, and feeder must be 40 % bigger to deliver the same useful work. Utilities pass this oversizing cost back through PF penalties: typically $3–$15 per kVA demand per month, or a kVA-based demand charge instead of kW. Correcting from 0.7 to 0.95 saves both the penalty and the demand charge premium.

Where do I install the capacitor bank?

Three options. (1) At the load — best for large steady motors; the capacitor cancels VAR right at the source. (2) At the motor control centre — corrects all motors on that MCC; common for industrial subdistribution. (3) At the main service / utility meter — corrects the billable PF but does not unload the internal feeders. Most installations combine (1) for the largest motors and (3) for the building average.

What is active power factor correction (active PFC)?

A switching converter (boost or other topology) that shapes the input current to follow the input voltage waveform — drawing nearly sinusoidal current at PF ≈ 0.99. Mandatory in modern power supplies above 75 W (EN 61000-3-2). Active PFC eliminates the need for line-side capacitor banks for those loads — datacenters with active-PFC PSUs typically run PF 0.95–0.98 without correction.

What if I have harmonics from VFDs?

Plain capacitor banks resonate with system inductance at some harmonic frequency, amplifying that harmonic current 3–10×. The cure is a detuned bank: each capacitor stage is in series with a small reactor that lowers the resonance frequency below all relevant harmonics (typically 134 or 189 Hz at 50/60 Hz line). For very high harmonic loads, use an active harmonic filter instead of capacitors.

Sources and methodology

1. IEEE. IEEE Std 18-2012 — IEEE Standard for Shunt Power Capacitors.
2. IEEE. IEEE Std 1036-2010 — IEEE Application Guide for Shunt Power Capacitors.
3. NEMA. NEMA CP 1 — Shunt Capacitors.
4. IEC. IEC 60831 — Shunt power capacitors of the self-healing type for AC systems with rated voltage up to and including 1 000 V.
5. IEC. IEC 61000-3-2 — Limits for harmonic current emissions. Active PFC mandate for equipment ≥ 75 W.
6. Schneider Electric. Electrical Installation Guide, latest edition. Chapter L — Power factor correction.
7. IEEE. IEEE Std 519 — Recommended Practice and Requirements for Harmonic Control in Electric Power Systems, 2014.