Power factor is the ratio of the average electrical power transferred by an alternating-current circuit to the product of its root-mean-square voltage and current. It describes how effectively current uses the power-carrying capacity of conductors, transformers, generators, and other equipment. A load with a low power factor requires more current than a high-power-factor load receiving the same average power at the same voltage.
Power factor is important throughout electrical engineering, from small electronic power supplies to industrial installations and the electrical grid. Low power factor can result from a phase difference between voltage and current, from distortion of their waveforms, or from both. These causes require different remedies: a capacitor bank can compensate the phase displacement associated with an inductive load, whereas distorted current generally requires filtering or electronic control.
The distinction between power factor and efficiency is fundamental. Efficiency compares useful output power with input power; power factor compares average electrical power transfer with the voltage–current capacity occupied by that transfer. An efficient motor can have a modest power factor, while a resistive heater can have nearly unity power factor without providing the same energy service efficiently.
Definition and physical meaning
For a single-phase circuit, instantaneous power is the product of instantaneous voltage and current:
The active power, also called real power, is its average over a suitable interval. For periodic steady-state operation with period ,
Active power is measured in watts and represents the net rate at which electrical energy is transferred. In a load, that energy may become mechanical work, heat, light, chemical energy, or losses. In a generating device, its direction can be reversed.
The root-mean-square values of voltage and current are
Their product is the single-phase apparent power,
measured in volt-amperes. Power factor is therefore
For a consuming load with positive active power, power factor lies between zero and one. If power flow is treated algebraically, the ratio can range from to . Instruments do not all use the same sign convention: a negative display may indicate reverse active-power flow or, under another convention, a leading rather than lagging condition. Its meaning depends on the meter’s definitions.
The mathematical bound follows from the Cauchy–Schwarz inequality. Equality requires current to be proportional to voltage throughout the averaging interval. For positive power flow, a constant positive conductance satisfies this condition: . Thus, even with distorted voltage, an ideal resistor has unity true power factor because its current follows the voltage waveform exactly.
A useful physical interpretation is that power factor measures current utilization. At fixed voltage and active power,
A 10 kW single-phase load supplied at 230 V draws approximately 43.5 A at unity power factor, but approximately 62.1 A at a power factor of 0.70. The additional current need not represent additional useful energy consumption. Nevertheless, it occupies conductor and equipment capacity and produces additional resistive losses.
Power factor is defined at a particular measurement boundary. A motor terminal, a factory incomer, and a utility substation can have different values because local loads and compensating devices combine differently at each location.
Sinusoidal circuits, phase angle, and reactive power
In sinusoidal steady state, voltage and current have the same frequency but may differ in phase. Writing
where and are RMS values, gives
Consequently, power factor is under these conditions. This familiar expression is not a general definition: it is valid when the relevant voltage and current waveforms are sinusoidal.
An inductive load draws current that lags voltage. An ideal inductor stores energy in a magnetic field during part of each cycle and returns it during another part. A capacitor stores energy in an electric field and draws current that leads voltage. Neither ideal component consumes average power, although both carry current. Practical inductors and capacitors also dissipate some active power through resistance, dielectric loss, and other mechanisms.
The conventional quantity describing this oscillatory exchange in a sinusoidal circuit is reactive power:
Reactive power is measured in var, or volt-amperes reactive. Under the usual passive-load convention, inductive loads have positive , while capacitive loads have negative . Active and reactive power can be represented together as complex power:
Its magnitude is apparent power, so that
This relationship is commonly drawn as a power triangle. Its horizontal side represents active power, its vertical side reactive power, and its hypotenuse apparent power.
For example, a sinusoidal load consuming 100 kW at 0.80 lagging power factor requires 125 kVA and absorbs 75 kvar. If a compensating device supplies 75 kvar locally, the upstream source ideally supplies only the 100 kW component, apart from compensation losses. The inductive equipment still requires its magnetizing current; compensation changes where the corresponding reactive current flows.
A power-factor number alone does not identify whether the load is inductive or capacitive. Since , both a leading and a lagging load can have a power factor of 0.90. The terms leading and lagging, or an explicitly defined reactive-power sign, are therefore necessary for a complete description.
The same distinction matters for zero-power devices. An ideal inductor or capacitor has zero power factor despite exchanging energy with the source every cycle. A practical unloaded transformer also commonly has low power factor: it draws magnetizing current while consuming only enough active power to cover its core and winding losses.
Three-phase systems and unbalanced loads
Most large power systems use three-phase electric power. For a balanced, sinusoidal three-phase load,
where is RMS line-to-line voltage and is RMS line current. These expressions apply to balanced star- or delta-connected loads when line quantities are used consistently.
A balanced 100 kW load on a 400 V supply draws approximately 180 A at a power factor of 0.80. Raising the power factor to 0.95 reduces its current to approximately 152 A. The active power is unchanged, but the reduction affects all three line conductors and the upstream transformer. For large installations, this difference can release a significant amount of distribution capacity.
Real installations are not always balanced. Single-phase lighting, office equipment, heating circuits, and domestic loads may be distributed unevenly among phases. Total active power remains the sum of the phase powers:
There is, however, no single universally interchangeable way to define total apparent power under all combinations of imbalance and waveform distortion. Adding individual phase apparent powers gives an arithmetic apparent power. Other definitions use effective voltage and current quantities to represent equivalent conductor utilization or losses. These definitions agree under balanced sinusoidal conditions but can differ appreciably elsewhere.
This distinction prevents a simple arithmetic average of three phase power factors from serving as a reliable system power factor. A lightly loaded phase with poor power factor should not receive the same weighting as a heavily loaded phase. Likewise, even a ratio based on the sum of phase apparent powers may not fully describe the additional capacity occupied by imbalance.
In a four-wire system, neutral current adds another complication. Balanced fundamental-frequency currents sum to zero, but certain harmonic currents do not. Third-harmonic currents from similar single-phase nonlinear loads are in phase across the three phases and can add in the neutral conductor. The same applies to other odd triplen harmonics, such as the ninth and fifteenth. A low fundamental displacement angle therefore does not guarantee low neutral current.
Measurement connections must also match the system. In a three-wire circuit, total active power can be measured by the two-wattmeter method. Under balanced sinusoidal conditions, the two readings can additionally be used to derive the phase angle. Under unbalance or distortion, their sum still provides total active power when the method is correctly applied, but the simple phase-angle interpretation no longer holds. Four-wire systems commonly use three measurement elements to capture each phase’s contribution.
Waveform distortion and nonlinear loads
A nonlinear load draws current that is not proportional to the instantaneous supply voltage. Diode rectifiers, electronic lighting drivers, computer power supplies, and many motor drives are examples. A bridge rectifier feeding a large DC-link capacitor often conducts mainly near the peaks of the AC voltage, producing narrow current pulses rather than a sine wave.
Such current can be analyzed into a fundamental-frequency component and harmonics at integer multiples of that frequency. Its fundamental component may be nearly aligned with the voltage, giving a high displacement power factor, while its overall RMS current is substantially increased by harmonics. Its true power factor, calculated from total active power and total RMS quantities, can therefore remain low.
For purely sinusoidal voltage and periodic current without a DC component,
and
Here is the RMS fundamental current, and is its phase displacement relative to the voltage. The first factor is the current distortion factor; the second is displacement power factor.
Current total harmonic distortion is commonly defined as
Under the same assumptions,
For example, with unity displacement power factor and current THD of 100%, true power factor is approximately 0.707. A meter reporting only would conceal the additional current burden.
When voltage is also distorted, the simplified factorization is generally insufficient. For periodic waveforms without DC components, average power includes contributions from matching voltage and current harmonics:
Different harmonic orders do not contribute cross terms to the average over their common period, but each matching order can transfer active power. Harmonic power can flow in different directions at different frequencies.
These conditions also limit the traditional power triangle. Defining produces a nonactive-power magnitude, but that quantity is not necessarily equivalent to fundamental reactive power and cannot generally be eliminated by adding capacitors. Distortion, imbalance, and phase displacement are physically different contributors to unused current capacity.
This is why a nonlinear power supply with low power factor should not automatically be treated like an induction motor with the same numerical power factor. A capacitor may correct the motor’s fundamental reactive demand, while doing little for the rectifier’s pulsed current and potentially creating a harmonic resonance.
Effects on networks, equipment, and electricity costs
The most direct consequence of low power factor is increased RMS current for a given active-power transfer. In a conductor of resistance , resistive heating is
For otherwise identical conditions, operating at 0.80 rather than unity power factor requires 25% more current and produces 56.25% more resistive loss in that conductor. Raising power factor from 0.80 to 0.95 reduces current to about 84.2% of its former value and reduces current-squared losses in the affected upstream section by about 29%.
These percentages describe the relevant network losses, not the entire installation’s electricity consumption. A factory receiving 100 kW does not reduce its production equipment’s active-power demand to 80 kW merely by correcting a power factor of 0.80. Savings arise from reduced losses, altered operating conditions, and tariff provisions. Compensation installed at a main switchboard also does not remove reactive current from motor feeders downstream of that board.
Equipment capacity is another important constraint. Transformers are commonly rated in kVA because winding current and applied voltage help determine their thermal and electromagnetic limits. A 1 MVA transformer supplying a load at 0.80 power factor transfers 800 kW; at 0.95 it transfers 950 kW at the same apparent-power loading. This comparison assumes that waveform distortion, cooling, and other operating limits remain acceptable.
Low lagging power factor can also increase voltage drop. In a simplified balanced feeder model, the voltage-drop component along the receiving-end voltage is related to
where and are feeder resistance and reactance. Locally supplied capacitive reactive power can reduce the reactive contribution. Excessive capacitive compensation, however, can instead raise voltage, particularly on lightly loaded feeders.
Generators face related limits: stator current constrains apparent power, while field excitation and machine heating constrain the permitted reactive-power range. A generator’s capability curve therefore contains more information than a single rated power-factor value.
Electricity tariffs may reflect these network costs through maximum kVA demand charges, reactive-energy charges, or penalties when power factor falls below a specified threshold. Other tariffs bill primarily for active energy and impose no direct power-factor charge. Metering intervals and contractual definitions matter, especially where demand varies sharply.
Power factor is also distinct from load factor. Load factor compares average demand with maximum demand over an interval. An installation can have excellent power factor but a poor load factor because it operates only briefly at high demand; conversely, a continuously operating plant can have a high load factor and substantial reactive demand.
Capacitor banks and passive compensation
For a sinusoidal inductive load, capacitors provide leading reactive current that offsets part of the lagging current drawn from the supply. The capacitor does not create active energy. Instead, it exchanges stored energy locally with inductive equipment, reducing the reactive current that must travel through the upstream network.
The required compensation for a load with active power , initial phase angle , and target phase angle is
For the 100 kW load initially operating at 0.80 lagging power factor, correction to 0.95 requires approximately 42.1 kvar. The initial reactive demand is 75 kvar, while the target demand is approximately 32.9 kvar.
A single-phase capacitor connected across RMS voltage supplies a reactive-power magnitude
where . For a balanced three-phase delta bank with capacitance in each branch,
For a star-connected bank with branch capacitance ,
Thus, an ideal 42.1 kvar delta bank on a 400 V, 50 Hz system requires approximately in each branch. Actual equipment selection also depends on voltage tolerance, harmonics, switching duty, temperature, and protection requirements.
Compensation may be individual, grouped, or centralized. A capacitor installed near a motor reduces reactive current through most of that motor’s supply path. A centralized bank is easier to coordinate across many varying loads but leaves reactive current flowing in downstream circuits. Automatic banks switch discrete capacitor steps according to measured demand, usually maintaining a target band rather than attempting exact unity power factor at every instant.
Capacitor switching creates engineering complications. Energization can produce high transient currents, particularly when another bank is already connected. Contactors, reactors, fuses, discharge resistors, and circuit breakers must accommodate the relevant duties. Capacitors connected with rotating induction machines can also support self-excitation under some disconnected or coasting conditions.
Network inductance and correction capacitance form a resonant system. If its resonant frequency lies near a significant harmonic, voltage or current distortion may be amplified. Detuned banks therefore include series reactors to shift the resonant condition, while tuned passive filters deliberately provide a low-impedance path for selected harmonics.
A further passive option is a line or DC-link reactor associated with a rectifier. It can broaden current conduction and reduce harmonic distortion, although it does not provide the same function as a shunt correction capacitor. Passive solutions must therefore be matched to whether the principal problem is phase displacement, current distortion, or both.
Active correction and power-electronic control
Electronic power-factor correction changes the current waveform through controlled switching. In an active front end, semiconductor devices regulate energy transfer so that the input current follows a desired reference. For a single-phase supply with approximately sinusoidal voltage, the usual objective is a sinusoidal current in phase with that voltage.
A common implementation is a boost-type correction stage placed after a diode bridge and before the main DC converter. A relatively slow voltage-control loop regulates the DC-link voltage, while a faster current-control loop shapes the rectified input current. The current reference is related to the rectified line voltage and the power required by the load.
The DC-link capacitor has an important energy-buffering role. Even at unity power factor, single-phase sinusoidal input produces instantaneous power that oscillates at twice line frequency:
for a suitable choice of phase origin. A downstream load consuming nearly constant power cannot receive that oscillating input directly without intermediate storage. Power-factor correction therefore does not eliminate the need for energy buffering; it changes how energy is drawn from the AC source.
Practical active correction must balance several objectives. Switching frequency affects magnetic-component size and switching losses. Input filters suppress high-frequency electromagnetic interference but can introduce capacitive current at light load. Current sensing, switching dead time, line-voltage distortion, and minimum conduction intervals can prevent perfect waveform tracking. Consequently, a unit that performs well near rated output may have a lower power factor at very light load.
Interleaved converters divide current among multiple switching channels, reducing ripple and spreading thermal stress. Bridgeless and totem-pole arrangements reduce some conduction losses associated with the conventional diode bridge. Three-phase active rectifiers can regulate DC-link voltage while controlling active and reactive current, and some permit bidirectional power flow.
These techniques are used in computer and server power supplies, charging equipment for the electric vehicle, and selected industrial drives. A variable-frequency drive with a conventional diode input bridge may already have a high fundamental displacement power factor while retaining appreciable current distortion. An active front end can address both, whereas adding a simple shunt capacitor mainly addresses fundamental reactive demand.
Active compensation can also be installed separately from the load. A shunt active power filter injects controlled currents that counter selected harmonic or reactive components. Its performance is constrained by converter current rating, control bandwidth, supply conditions, and the accuracy of the load-current measurements. A device sized for reactive compensation alone may lack sufficient capacity to cancel large harmonic currents simultaneously.
Measurement, reporting, and standards
Reliable power-factor measurement requires simultaneous knowledge of voltage, current, and average power. A modern power analyzer samples voltage and current, multiplies corresponding samples to determine active power, calculates true RMS values, and forms their ratio. This approach can capture distorted waveforms provided that the sensors, sampling rate, bandwidth, and numerical processing are suitable.
An instrument that assumes sinusoidal current may give misleading results on electronic loads. Average-responding meters calibrated in RMS units are accurate only for the waveform assumptions built into their calibration. Even true-RMS instruments have limits, including frequency bandwidth and crest factor—the ratio of peak to RMS magnitude. Narrow rectifier-current pulses can exceed an input channel’s peak capability before its displayed RMS range is exceeded.
Phase accuracy is especially important at low power factor. If voltage and current are nearly 90 degrees apart, a small sensor phase error can represent a large fraction of the measured active power. Current transformers, flexible current probes, voltage dividers, and acquisition channels can each introduce phase shifts. Harmonic measurements further require that these errors remain controlled over the frequencies of interest.
Reports should distinguish true power factor from fundamental displacement power factor. A display marked “cos phi” often refers to fundamental phase displacement rather than the ratio of total active to apparent power, although instrument documentation determines the exact interpretation. Leading or lagging status, active-power direction, phase configuration, and the apparent-power definition are also relevant.
Time aggregation introduces another distinction. An instantaneous operating-point value, an average of short-interval power factors, and an interval value calculated from accumulated quantities need not be equal. In particular, averaging power-factor numbers without weighting them by the associated load can produce an unrepresentative result. Utility billing registers may use definitions tailored to demand or reactive-energy accounting.
IEEE Standard 1459 provides definitions for power quantities under sinusoidal, nonsinusoidal, balanced, and unbalanced conditions. Its importance is not that every meter must display one universal power-factor number, but that different contributions and definitions can be stated explicitly rather than conflated.
The IEC 61000 series addresses related aspects of power quality. IEC 61000-3-2 concerns harmonic-current emissions from equipment with rated input current up to and including 16 A per phase, within its specified scope. IEC 61000-3-12 addresses equipment above 16 A and up to 75 A per phase connected to public low-voltage systems. These are harmonic-emission standards, not blanket requirements that every device achieve one specified power factor.
IEEE Standard 519 addresses harmonic control at the point of common coupling. Its current-distortion framework includes total demand distortion, which uses a demand-current reference rather than the instantaneous fundamental current used in ordinary current THD. A high power-factor reading alone therefore does not establish compliance with harmonic limits.
Motors, distributed generation, and grid operation
The behavior of induction motors explains many traditional industrial power-factor problems. A motor needs magnetizing current to establish its rotating magnetic field. This component does not decline in proportion to mechanical output, so an unloaded or lightly loaded motor commonly has a lower power factor than the same motor operating near rated load.
As shaft load increases, the active-current component grows relative to the magnetizing component, usually improving power factor over much of the normal operating range. Motor efficiency and power factor nevertheless remain separate characteristics. Mechanical losses, winding losses, and core losses determine efficiency, while the relative active, reactive, and distorted current components determine power factor.
Synchronous machines offer a different form of control. By adjusting excitation, a synchronous motor or generator can absorb or supply reactive power within its operating limits. A synchronous condenser is a synchronous machine operated primarily for reactive-power and voltage support rather than mechanical output. Unlike a capacitor bank, it is a rotating device; it also contributes inertia and fault current in ways that power-electronic compensators do not inherently reproduce.
In solar-energy and wind-power installations, a power inverter can often control reactive current independently of active current within its rating. Unity-power-factor operation maximizes active output for a fixed current or apparent-power limit, but grid operation may call for reactive support instead. A converter operating at its apparent-power limit must reduce active output if additional reactive output is required, unless it has unused capacity.
For an idealized converter with apparent-power limit ,
Real capability envelopes are additionally affected by AC voltage, DC-link conditions, semiconductor temperatures, and control limits. Near zero active output, power factor becomes a particularly poor description of service: a device can provide substantial reactive support while having a power factor near zero.
Distribution systems with extensive renewable energy may use fixed reactive-power settings, fixed power-factor settings, or voltage-dependent reactive-power control. The latter varies reactive output in response to measured voltage. On resistive low-voltage feeders, active-power flow can also have a strong effect on voltage, so reactive control is not a universal substitute for other network measures.
The appropriate power factor therefore depends on location and operating purpose. An industrial customer may seek to reduce imported reactive power, while a network operator may deliberately request reactive absorption or injection from that same site. During export, the active-power direction reverses, making an explicit sign convention essential.
In microgrids and converter-dominated networks, power factor is only one operating indicator. Voltage stability, harmonic sharing, phase imbalance, current limits, and the coordination of grid-forming and grid-following controls can remain important even when aggregate power factor is close to unity.