f97.beERIK SCHMIDT’S WORKSHOP
← All terms

Glossary · Energy industry & energy transition

Reactive power

Reactive power voltage control grid operation alternating current converter transformer ancillary service electricity grid

Reactive power is the portion of electrical power in an alternating current system that oscillates back and forth between electric or magnetic fields and the grid. On average, it is not converted into heat, motion, light or chemical energy. Unlike active power, it performs no useful work. Nevertheless, it causes currents to flow, influences the voltage and loads lines, transformers and switchgear. In the power grid, reactive power is therefore not a side issue but a key operating variable for voltage control.

The symbol is Q, the unit is var (volt-ampere reactive). In practice, it is usually expressed in kvar or Mvar. Together with the active power P (in watts), it yields the apparent power S (in volt-amperes). For sinusoidal quantities, S² = P² + Q² applies, and also P = S · cos φ and Q = S · sin φ. Here, φ is the angle by which the current is shifted relative to the voltage (see phase position and phase angle). The cosine of this angle is called the displacement factor cos φ. Only for purely sinusoidal voltages and currents is it equal to the power factor λ = P/S. When harmonics occur, for example due to power electronics, the power factor is smaller than cos φ. The two terms should therefore not simply be equated.

A worked example: An inverter with 10 kVA apparent power operates at cos φ = 0.90. It can then deliver at most 9 kW of active power and at the same time provides about 4.4 kvar of reactive power. At cos φ = 1, it could feed in the full 10 kW.

Physical cause: inductive and capacitive

The cause lies in the alternating voltage itself. Coils, motors, chokes and transformers build up and break down a magnetic field in every half-cycle. In doing so, the current lags behind the voltage. Such equipment draws inductive reactive power. For capacitors, capacitor banks and cables, it is the other way round: the electric field causes the current to lead the voltage. This equipment supplies reactive power, thus acting capacitively. How inductance and capacitance behave towards the grid depends only on the phase position. A clear direction only results once a sign convention (reference-arrow system) is defined. In grid connection rules, generating plants are therefore usually described as operating underexcited (the plant draws reactive power and lowers the voltage) or overexcited (it supplies reactive power and raises the voltage). The terms originate from the synchronous machine. Anyone reading specifications such as “cos φ 0.95 inductive” should therefore always check from whose perspective the sign is counted.

The distinction between reactive power and losses is important. In the idealised case, reactive power oscillates back without loss. However, the additional current it causes heats lines and transformers and thus increases grid losses. Equipment is designed for current and voltage, that is, for apparent power, not just for the active power transmitted. A transformer with a fixed rated apparent power therefore transmits less active power when a large amount of reactive power flows at the same time.

Reactive energy and billing

Reactive power must be distinguished from reactive energy, that is, reactive power over a period of time, measured in varh, kvarh or Mvarh. For commercial and industrial customers with demand metering, reactive energy can be billed separately if it exceeds a free allowance set out in the grid operator’s price sheet. Whether and at what level this happens is determined by each grid operator itself. Households generally pay only for active energy in kilowatt-hours. Industrial companies usually reduce such costs through reactive power compensation, for example with capacitor banks close to large motors.

Why reactive power determines voltage

In high-voltage and extra-high-voltage grids, line impedances are predominantly inductive. How much the voltage drops or rises over a given distance depends there mainly on the reactive power flow. The further one goes down into medium- and low-voltage grids, the more active power and ohmic resistance also play a role. This has practical consequences for grid operators: a photovoltaic system at the end of a long low-voltage feeder raises the voltage when feeding in at high output. If it draws reactive power in underexcited operation at the same time, it can partly offset this rise. A large motor start, by contrast, lowers the voltage.

Reactive power acts locally. It can be transported over long distances, but this causes additional currents, losses and voltage drops. The EU grid connection regulation expressly states that the need for reactive power capability depends, among other things, on the degree of meshing of the grid and on the ratio between generation and consumption. A voltage problem in the distribution grid can therefore generally not be solved by a distant plant in the transmission grid. Often the location matters more than the sheer quantity in Mvar.

Who provides reactive power

In the past, a large share of controllable reactive power came from synchronous generators in large power plants. Through their excitation, they can absorb or supply reactive power. When conventional power plants run less often, the demand remains, but familiar sources drop out. They are replaced by:

  • Converters of wind, photovoltaic and battery storage systems. An inverter can provide reactive power within its apparent power limit. At full active power, however, only little headroom remains for this.

  • STATCOM systems, SVCs, capacitor banks and compensation reactors at substations.

  • Rotating phase shifters: synchronous machines without a drive, which additionally provide short-circuit power and instantaneous reserve.

  • Transformers with an tap changer, which do not generate reactive power themselves but change the voltage level and thus influence reactive power flows.

How grid operators coordinate these sources is described by reactive power management. Several control methods exist for this: fixed cos φ, characteristic curves such as cos φ(P) or Q(U), and remotely controllable setpoints.

Legal framework in the EU and Germany

In the EU, Regulation (EU) 2016/631 (Requirements for Generators, RfG) lays down minimum requirements for power-generating facilities. The requirements are graded by plant types A to D. For non-synchronous plants of type C, Article 21 contains, among other things, provisions on reactive power capability as well as on voltage, reactive power and power factor control. The exact ranges are specified by the respective competent grid operator within the framework of the regulation.

In Germany, the technical connection rules of VDE FNN implement these requirements, separated by voltage level: VDE-AR-N 4105 for low voltage, 4110 for medium voltage, 4120 for high voltage and 4130 for extra-high voltage. In addition, there are the technical connection conditions of the grid operators. According to an FNN guidance note from 2014, VDE-AR-N 4105 requires generating plants in the low-voltage grid, depending on their size, to provide a displacement factor of 0.95 or 0.90 underexcited to overexcited. Since 1 January 2012, these plants have been delivered with the cos φ(P) characteristic curve activated. The exact power thresholds depend on the valid version of the standard. Which setting applies in an individual case is specified by the connecting grid operator.

Since 27 November 2020, Section 12h EnWG has in principle obliged grid operators to procure the non-frequency ancillary service “voltage control services” in a transparent, non-discriminatory and market-based manner. With decision BK6-23-072 of 25 June 2024, the Federal Network Agency (Bundesnetzagentur) specified the details. The determination is addressed to the control-area-responsible transmission system operators and to distribution system operators with a high-voltage grid. Operators of low- and medium-voltage grids are exempt from the obligation. The authority justifies this by the low market volume and high transaction costs. Reactive power capabilities that are already mandatory under TAR or TAB are not procured through the market. If an obligated grid operator has a need, it must initiate a procedure for at least one procurement region no later than twelve months after the determination. TenneT announced that the transmission system operators will introduce the reactive power market for the first time in 2025.

Typical misinterpretations

  • “Reactive power is unnecessary current”: Reactive power is not an additionally consumed kilowatt-hour. However, it causes additional current, which generates losses and occupies transmission capacity. At the same time, it is indispensable for maintaining voltage.

  • “Reactive power can be traded centrally like quantities of electricity”: Because it is so closely coupled to the voltage and the grid impedances, it acts within a limited area. Market-based procurement is therefore also organised by region.

  • “Reactive power maintains the frequency”: The frequency depends mainly on the balance between generated and consumed active power, see frequency control. Reactive power acts primarily on the voltage. The two are linked in operation, but they are different tasks.

  • “cos φ and power factor are the same thing”: This is only true for purely sinusoidal quantities. With harmonics, the two values differ from each other.

Reactive power shows that a power grid cannot be understood through energy quantities alone. The grid must keep voltages within the permissible band, limit currents and control electrical conditions locally. Reactive power is not a usable quantity of electricity. It is a prerequisite for active power to be transmitted reliably. In a power system with ever more converters, how well this succeeds increasingly depends on connection rules, control engineering and procurement procedures.

Sources

Related glossary entries

Edited by eCMS on 11.11.2026 16:08

© 2026 Erik Schmidt. All rights reserved. Quotations with a link to the source are welcome.