The formula
Q_c = P × (tan φ₁ − tan φ₂), where φ = arccos(pf)
- P
- real power of the load, unchanged by correction (watts)
- pf₁, pf₂
- power factor before and after correction
- Q_c
- reactive power the capacitor must supply (VAr)
Worked example
A 10 kW load at 0.7 lagging, to be corrected to 0.95.
- φ₁ = arccos(0.7) = 45.6°, so tan φ₁ = 1.020
- φ₂ = arccos(0.95) = 18.2°, so tan φ₂ = 0.329
- Q_c = 10 × (1.020 − 0.329)
Q_c ≈ 6.9 kVAr of capacitance, which drops the line current by about 26 %.
Where you'll use it
Industrial energy audits and the standard exam question on capacitor bank sizing. Correcting all the way to unity is rarely done — the last few percent needs disproportionately large capacitors and risks leading power factor at light load.