Summary Highlights
- Correction can supply part of a load’s reactive demand locally and reduce current upstream of its connection.
- Total power factor and fundamental displacement cosφ can differ when waveforms contain harmonics.
- Location, switching control and reactor design are separate choices, not competing names for one classification.
- A kvar calculation needs compatible measurements and a representative load profile.
- Billing, transformer loading and harmonic performance require separate evidence.
Article Details
Power factor correction changes the reactive-power balance of an installation to improve its power factor. For an inductive load, a common method is to provide some reactive power locally with a capacitor bank. The supply upstream of that connection can then carry less reactive current for the same active load. The outcome depends on the connection point, load variation, waveform distortion and control system; installing capacitors alone does not establish a saving or an acceptable design.
Active, reactive and apparent power
Active power P, in kW, represents electrical power converted into work or heat. Reactive power Q, in kvar, describes an exchange associated with electric and magnetic fields. Apparent power S is expressed in kVA. An induction motor still needs its magnetic field after correction: the bank changes where part of its reactive demand is supplied rather than removing that physical requirement.
The familiar power triangle is useful under its sinusoidal assumptions. With distorted waveforms, it does not by itself explain every total-power quantity. Meter definitions, phase balance and the direction of power flow must also be considered.
Is power factor the same as cosφ or efficiency?
Power factor is PF = P/S. It is different from a machine’s energy-conversion efficiency. Displacement power factor describes the phase relation between fundamental voltage and current components. Under sinusoidal conditions it can equal total PF; with significant harmonic current, displacement cosφ can remain close to unity while total PF is lower.
| Reading | What it describes | What it does not establish |
|---|---|---|
| Total PF | Active power divided by the defined apparent power. | Conversion efficiency or a uniquely inductive cause of low PF. |
| Displacement cosφ / DPF | The fundamental voltage–current phase relationship. | Low harmonic distortion or an identical total PF. |
| THD and harmonic spectrum | Distortion of the identified voltage or current. | A complete bank or filter selection from one percentage. |
| Reactive-energy ratio | The relationship of particular energy registers over a stated interval. | The bill from one instantaneous PF reading. |
How capacitor banks and automatic controllers work
An automatic controller uses its voltage and current measurements to select available capacitor steps. The CT location, measured phase, ratio settings, step sizes and switching delays all affect the decision. A measurement error or an oversized first step can produce overcompensation or excessive switching even when the capacitors are sound.
Contactor-switched and thyristor-switched banks have different operating characteristics. Faster-changing loads may justify a suitable electronic switching design, but response time and permitted duty belong to the actual equipment specification. An SVG is a power-electronic compensation device, not simply a capacitor contactor with another name. Active PFC in an AC/DC power supply is also a different application from an industrial automatic bank.
Fixed, automatic, local, group and central correction
| Design choice | Meaning | Condition to examine |
|---|---|---|
| Fixed or automatic | A defined amount of support, or steps managed as the load changes. | Minimum load, step size, switching duty and available controls. |
| Local / individual | Correction close to a load requiring reactive power. | Load interlocking, motor self-excitation risk and manufacturer connection limits. |
| Group | A common bank serving a group of loads. | Whether the loads actually operate together and which feeder section benefits. |
| Central | A common bank at a main distribution point. | Downstream load cables can still carry their reactive current. |
| Detuned or combined with filtering | An impedance and compensation arrangement selected for the harmonic environment. | Network impedance, harmonic sources and matched equipment ratings. |
For example, a central installation can be both automatic and detuned. These labels describe different features. A fixed bank need not be assumed to remain connected during every shutdown; the actual switching scheme governs. A capacitor must not be connected arbitrarily to a motor or drive output without the equipment-specific design requirements.
How much correction is needed?
A preliminary sinusoidal calculation is Qc = P × (tanφ1 − tanφ2), using active power P and compatible initial and target displacement angles. Substituting a distortion-affected total PF as though it were cosφ can give an inappropriate result. Final selection also needs representative high-load and low-load conditions.
Consider an illustrative balanced, sinusoidal three-phase load at 400 V: P = 100 kW, initial cosφ = 0.80 and a chosen target of 0.95. The calculation gives about 42.1 kvar. Idealised supply current falls from about 180.4 A to 151.9 A, while the active load remains 100 kW. These are explanatory assumptions, not Pow-Sys measurements, a universal target or an order specification.
The bank must be considered with existing support, the smallest usable step, switching frequency, load change rate, voltage and harmonic conditions. A plain capacitor’s kvar varies with frequency and the square of its applied voltage for unchanged capacitance. A detuned bank’s net kvar at the supply and a capacitor element’s nameplate kvar at its rated voltage are not automatically the same quantity.
Does a detuned bank remove harmonics?
A capacitor and the inductive network can create a resonant condition. A detuned branch uses a series reactor and capacitor selected together for a particular tuning relationship. It does not remove every harmonic, nor guarantee that resonance is impossible for every network and switching state.
What changes on the electricity bill or transformer?
A suitable correction system may reduce reactive-energy or low-PF charges where the applicable tariff imposes them. That is separate from an active-kWh saving. Some upstream losses can decrease, while the bank and its auxiliaries also have losses. The meter boundary, load profile and actual charging rules determine the economic result; a tariff example from another country is not transferable by default.
When a relevant ratio is exceeded, the measured reactive amount is considered in full, rather than only the excess above the limit. If both directions exceed their limits, the larger reactive meter record is considered in full, rather than automatically summing both. The first breach in a calendar year, determined by the bill’s final reading date, is shown and deducted on the same bill. A reading longer than 35 days also has a 35-day charge cap. Reactive export while a producer supplies active energy has a separate provision and Annex 1 methodology; import and export must not be casually netted. Both reactive records being zero is separately addressed in Article 13(7), not an automatic exemption. An exact bill still requires the applicable tariff, scope and actual registers; no current unit price or guaranteed reduction is asserted here.
Reducing upstream current can release part of an existing cable or transformer kVA allowance for active load. It does not increase a transformer’s nameplate rating. Harmonic heating, voltage, thermal limits and the real duty continue to constrain usable capacity.
Information to prepare for an assessment
Useful inputs include the single-line diagram, connection and installed-power information, active and reactive energy records, operating schedules and existing bank labels. Controller model and settings, CT details, step sizes, capacitor/reactor ratings and recurring fault history help define a meaningful measurement campaign.
Share the facility location, the observed problem and available documents with Pow-Sys to define an engineering or power-quality assessment. The site-visit scope and timing are confirmed together. One bill or one PF screenshot cannot establish a complete bank design or a guaranteed saving.
Related Blog Posts
Related Services
Frequently Asked Questions
Does power factor correction reduce kWh consumption?
It may reduce some upstream losses, but avoiding a reactive charge is not the same as reducing active kWh. Loading, the metering boundary, tariff rules and the bank’s own losses determine the result.
Why can total PF be low when cosφ is good?
Harmonic currents can increase apparent power while the fundamental phase relationship remains favourable. Review voltage and current waveforms and the meter’s definitions rather than treating cosφ and total PF as interchangeable.
Are fixed and central correction the same?
No. Fixed or automatic describes control; local, group or central describes position. A central bank can also be automatic and detuned.
Does a central bank reduce reactive current in every cable?
It can reduce current upstream of its connection. The downstream cables feeding an inductive load can still carry that load’s reactive current. The connection and metering points matter.
Should an APFC controller always target 1.00?
There is no universal target. Minimum load, step size, delays, harmonics, operational needs and applicable charging rules must be considered. Forcing unity can encourage unwanted capacitive operation.
Does adding a reactor solve every harmonic problem?
No. The reactor and capacitor form a selected impedance. Harmonic-filtering performance and resonance under different network and switching conditions require their own assessment.
Does negative PF always mean leading operation?
No. Instruments can use different sign conventions. Read active-power direction, reactive sign, connection and any generation condition using the actual meter definition.
Can transformer kVA alone determine bank size?
No. The active and reactive load profile, minimum demand, existing support, step arrangement, harmonic environment and equipment ratings are also needed.