Power Factor Correction and Harmonic Filtering: Capacitor Banks, Detuning, and Active Filters

Published: June 2026
Technical Level: Advanced Category: Power Quality

Abstract

Power factor correction and harmonic filtering are closely coupled problems in facilities that combine large motor loads with nonlinear electronic equipment, because the capacitor banks installed to correct displacement power factor can resonate with system inductance at harmonic frequencies and fail destructively if they are not detuned. This paper develops the engineering basis for the combined problem, distinguishing displacement power factor from true power factor in the presence of harmonics, deriving the reactive compensation required to reach a target power factor, and quantifying the system benefits of correction in reduced losses and released capacity. It then addresses the resonance hazard that untuned capacitors introduce, the detuning reactor that eliminates it, and the selection between passive tuned filters and active harmonic filters for facilities whose harmonic distortion must be reduced as well as whose power factor must be corrected. The objective throughout is to size correction and filtering equipment so that the facility meets both its utility power factor obligation and the harmonic limits of IEEE 519-2022 without creating a resonance condition.

1. Introduction

A facility's power factor measures how effectively it converts the apparent power drawn from the utility into useful real power. Induction motors, transformers, and other inductive loads draw a magnetizing reactive current that does no real work but that nonetheless loads the conductors, transformers, and utility supply that serve the facility. A low power factor therefore wastes system capacity, increases resistive losses, and, in most utility tariffs, incurs a direct financial penalty assessed on the reactive demand or on the power factor itself. The conventional remedy is to install shunt power-factor-correction capacitors that supply the reactive current locally, relieving the upstream system of the burden of carrying it.

The complication in a modern facility is that the same loads that depress power factor are frequently accompanied by nonlinear loads that inject harmonic currents, and a power-factor capacitor is not a benign addition to such a system. The capacitor forms a parallel resonant circuit with the inductance of the source and transformer, and if the resonant frequency of that circuit falls near a harmonic frequency present on the system, the capacitor amplifies rather than absorbs the harmonic current, with consequences ranging from nuisance fuse operation to the destruction of the capacitor itself. Power factor correction and harmonic control must therefore be designed together: the reactive compensation must be sized to meet the power factor target, and it must be detuned or filtered so that it does not resonate. This paper treats the two as the single coupled problem they are in practice.

2. Power Factor in the Presence of Harmonics

2.1 Displacement and True Power Factor

For a purely sinusoidal system, the power factor is the cosine of the phase angle between the voltage and current, a quantity known as the displacement power factor:

$PF_{disp} = \frac{P}{S_1} = \cos\phi$

Where:

$PF_{disp}$ is the displacement power factor.

$P$ is the real power in kilowatts.

$S_1$ is the fundamental-frequency apparent power in kilovolt-amperes.

$\phi$ is the phase angle between the fundamental voltage and current.

When harmonic currents are present, the total root-mean-square current exceeds the fundamental current, so the true apparent power exceeds the fundamental apparent power even though the harmonics carry no real power. The true power factor, which is the quantity a utility revenue meter actually measures, is the displacement power factor reduced by a distortion factor that depends on the total harmonic current distortion:

$PF_{true} = PF_{disp} \cdot \frac{1}{\sqrt{1 + \text{THD}_I^2}}$

Where:

$PF_{true}$ is the true power factor.

$PF_{disp}$ is the displacement power factor.

$\text{THD}_I$ is the total harmonic current distortion expressed as a decimal fraction.

The practical significance of this relationship is that capacitors correct only the displacement component. A facility with a displacement power factor of 0.78 and a current distortion of 18 percent has a true power factor near 0.77; installing capacitors can raise the displacement factor to 0.95, but the residual distortion continues to depress the true power factor unless the harmonics themselves are reduced. Capacitor correction alone cannot bring a harmonically distorted facility to a high true power factor, and a facility that is penalized on true power factor must address both the reactive demand and the distortion.

2.2 Sizing the Reactive Compensation

The capacitive reactive power required to raise the power factor from its present value to a target value is the difference between the reactive power consumed at the two power factors for the same real load:

$Q_C = P\left(\tan\phi_1 - \tan\phi_2\right)$

Where:

$Q_C$ is the required capacitor reactive power in kilovars.

$P$ is the real power in kilowatts.

$\phi_1$ is the phase angle corresponding to the existing power factor.

$\phi_2$ is the phase angle corresponding to the target power factor.

A facility drawing 1,200 kW at a displacement power factor of 0.78 has a phase angle $\phi_1 = \arccos(0.78)$ with $\tan\phi_1 = 0.802$ ; correcting to a target of 0.95 gives $\phi_2 = \arccos(0.95)$ with $\tan\phi_2 = 0.329$ . Substituting into the sizing equation:

$Q_C = 1200\,(0.802 - 0.329) = 1200 \times 0.473 = 568 \ \text{kvar}$

so a 568 kvar capacitor bank (rounded up to a standard 575 kvar bank) raises the displacement power factor from 0.78 to 0.95. The target power factor should be selected with care: correcting to unity, or beyond, risks a leading power factor and overvoltage under light-load conditions, particularly when the capacitors are fixed rather than automatically switched. A target in the range of 0.95 to 0.98 satisfies typical utility thresholds while preserving a margin against overcorrection, and automatic switched banks are preferred where the load and therefore the optimal compensation vary significantly through the operating day.

2.3 Loss Reduction and Released Capacity

Correcting power factor reduces the current that the conductors and transformers carry for the same real load, and because resistive losses vary with the square of the current, the loss reduction is substantial. The fractional reduction in resistive losses achieved by raising the power factor is:

$\Delta P_{loss} = 1 - \left(\frac{PF_{old}}{PF_{new}}\right)^2$

Where:

$\Delta P_{loss}$ is the fractional reduction in resistive losses.

$PF_{old}$ is the power factor before correction.

$PF_{new}$ is the power factor after correction.

Raising the power factor from 0.78 to 0.95 by this relationship reduces resistive losses by approximately 32 percent in the affected conductors and transformers. The same reduction in current also releases system capacity: the apparent power drawn for a fixed real load falls in proportion to the improvement in power factor, freeing transformer and feeder capacity that would otherwise have to be added to serve load growth. These two benefits, together with the avoidance of the utility power factor penalty, generally make power factor correction one of the most rapidly recovered investments in facility electrical engineering.

3. Resonance and Detuning

The hazard that distinguishes power factor correction in a harmonic environment from the simple sinusoidal case is parallel resonance. The shunt capacitor and the inductance of the source and supply transformer form a parallel resonant circuit whose resonant harmonic order is approximately:

$h_r = \sqrt{\frac{S_{sc}}{Q_C}}$

Where:

$h_r$ is the resonant harmonic order.

$S_{sc}$ is the short-circuit apparent power at the capacitor bus in megavolt-amperes.

$Q_C$ is the capacitor reactive power in megavars.

When this resonant order falls near a harmonic that the facility's nonlinear loads produce (most commonly the fifth, at 300 hertz on a 60 hertz system), the parallel circuit presents a high impedance to that harmonic and amplifies the resulting harmonic voltage and the current circulating between the source and the capacitor. The amplified current overheats and ultimately destroys the capacitor, and the elevated harmonic voltage stresses other equipment on the bus. This mechanism is the cause of the great majority of unexplained capacitor failures in industrial facilities.

The remedy is to add a series reactor to the capacitor, forming a detuned bank whose series-resonant frequency is placed below the lowest significant harmonic. A reactor sized to tune the bank to approximately 189 hertz blocks amplification at the third harmonic, while tuning to approximately 210 hertz, between the fundamental and the fifth harmonic, protects against fifth-harmonic resonance. Below its tuning frequency the detuned bank behaves as a net capacitance and delivers the intended power factor correction; above it the bank is inductive and cannot resonate with the source at any harmonic of concern. A detuned bank therefore provides power factor correction while remaining immune to the resonance that destroys an untuned bank, and detuning is the standard practice wherever capacitors are installed on a bus that carries appreciable harmonic current.

4. Harmonic Filtering

Where the facility must reduce its harmonic distortion as well as correct its power factor, to comply with the current limits of IEEE 519-2022 at the point of common coupling, the detuned bank is extended into a tuned passive filter or supplemented by an active filter.

A passive tuned filter is a series inductor-capacitor branch tuned at, rather than below, a target harmonic, presenting a low impedance that shunts that harmonic current away from the source while simultaneously providing reactive compensation at the fundamental. A filter tuned near the fifth harmonic addresses the dominant order produced by six-pulse rectifiers and variable-frequency drives, and a second branch near the seventh extends the coverage. The passive filter is robust and economical, and its dual function as power factor correction makes it efficient where the dominant harmonics are few and stable. Its limitations are that each branch addresses only one harmonic, that its performance depends on component tolerances and on the system impedance, and that it must be analyzed for resonance against the harmonics it does not target.

An active harmonic filter measures the load's harmonic current and injects an equal and opposite current, cancelling many harmonic orders simultaneously and adapting in real time to changing load. It introduces no resonance and can additionally correct power factor within its rating. The active filter is sized by the harmonic current it must compensate rather than by the load's total current; a rating of roughly 30 to 50 percent of the nonlinear load current is typical for facilities with moderate distortion. The selection between passive and active filtering follows the breadth and stability of the harmonic spectrum: a facility dominated by a few stable harmonic orders is well served by detuned or tuned passive equipment, while a facility with a broad or time-varying spectrum, or one where resonance risk is a concern, is better served by an active filter, often in a hybrid arrangement in which a passive branch handles the bulk fifth-harmonic load and a smaller active filter trims the residual and higher-order content.

The spectrum that motivates the filter design is shown in Figure 1, which plots the harmonic current magnitudes of the nonlinear load before mitigation.

Harmonic current spectrum of the uncorrected nonlinear load. The horizontal axis is harmonic order and the vertical axis is magnitude as a percentage of the fundamental. The dominant 5th and 7th harmonics establish the tuning frequencies.

Figure 1. Harmonic current spectrum of the uncorrected nonlinear load. The horizontal axis is harmonic order and the vertical axis is magnitude as a percentage of the fundamental. The dominant 5th and 7th harmonics establish the tuning frequencies for a detuned or tuned filter; the engineer should observe that a single tuned branch at the 5th harmonic captures the largest distortion component, while full compliance with IEEE 519 limits may require additional branches at the 7th and 11th orders.

5. Conclusion

The most consequential finding is that power-factor correction and harmonic filtering are a single coupled problem in any facility with both reactive motor load and nonlinear electronic load, because the capacitor bank that corrects power factor becomes a resonant hazard the moment it shares a bus with harmonic current. Treating the two as separate scopes is the error that turns a routine correction into an equipment-damaging resonance.

The most common implementation failure is installing an untuned capacitor bank on a bus carrying harmonics, which will eventually resonate near a system harmonic order and fail; the standard remedy, a series detuning reactor, is omitted precisely because the steady-state power-factor calculation that justified the bank never examined the harmonic spectrum it would interact with.

The engineer should next compute the resonant frequency of the proposed bank against the system inductance before energizing it, and extend the detuned bank into a tuned passive filter where IEEE 519-2022 limits must also be met, because the next problem the correction sets up, the resonance it can create, is the one that determines whether the installation operates or destroys itself.

Related Work

The analysis in this paper connects to several companion studies in this library. Readers concerned with the upstream and downstream engineering will find Harmonic Distortion in Commercial and Industrial Power Systems develops a closely related aspect of the same problem, while Power Quality Analysis extends the treatment into an adjacent domain. For the broader methodological context, Variable Frequency Drive Harmonic Mitigation in Manufacturing provides complementary depth.

References

[1] IEEE Standard 519-2022, IEEE Standard for Harmonic Control in Electric Power Systems, IEEE, 2022.

[2] IEEE Standard 18-2012, IEEE Standard for Shunt Power Capacitors, IEEE, 2012.

[3] IEEE Standard 1531-2003, IEEE Guide for Application and Specification of Harmonic Filters, IEEE, 2003.

[4] IEEE Standard 141-1993, IEEE Recommended Practice for Electric Power Distribution for Industrial Plants (Red Book), IEEE, 1993.

[5] NFPA 70-2023, National Electrical Code, National Fire Protection Association, 2023.

[6] R. C. Dugan, M. F. McGranaghan, S. Santoso, and H. W. Beaty, Electrical Power Systems Quality, 3rd ed., McGraw-Hill, 2012.

[7] J. Arrillaga and N. R. Watson, Power System Harmonics, 2nd ed., Wiley, 2003.

[8] T. J. E. Miller, Reactive Power Control in Electric Systems, Wiley, 1982.