Reactive power and voltage support from PV inverters

A PV inverter's most obvious job is converting DC array output into AC power at the correct voltage and frequency. A less obvious but increasingly mandatory job is helping the local network hold its voltage within limits, using a capability inverters have but PV panels themselves do not: the ability to shift the phase relationship between voltage and current and export or absorb reactive power. This article explains what reactive power support actually does to a distribution circuit, the main control modes grid codes require, and how to work through a simple voltage-rise calculation by hand.

By the end you should be able to read an inverter datasheet's reactive power capability curve, understand the difference between fixed power factor, cos-phi(P), and Q(U) droop modes, and recognise the common configuration mistakes that either waste inverter capacity or fail to deliver the voltage support a network operator is relying on.

Why voltage rise happens

On a distribution feeder, voltage naturally falls as current flows away from the substation toward loads, because of the resistance and reactance of the cable or overhead line. A PV installation exporting power reverses part or all of that current flow during sunny periods, which can push voltage upward rather than downward at the point of connection, particularly on long rural low-voltage feeders with thin conductors and several PV systems clustered together. If enough PV pushes local voltage above the statutory upper limit, network operators historically had to respond with network reinforcement — new conductors, a new transformer tap, or a voltage regulator — which is slow and expensive.

Reactive power exchange offers a faster, software-based alternative. Absorbing reactive power at the inverter (behaving inductively) pulls local voltage down slightly; exporting reactive power (behaving capacitively) pushes it up. On feeders dominated by resistance rather than reactance — common at low voltage — the effect of active power curtailment on voltage is actually stronger per kW than reactive power exchange per kvar, which is worth knowing before assuming reactive power support alone will always solve a voltage problem; but on stiffer or more reactive feeders, particularly medium voltage, reactive power support is genuinely powerful and far less wasteful than curtailing active power.

Control modes

Most European low-voltage connections use either a cos-phi(P) characteristic or Q(U) droop, both specified in EN 50549-1, with the exact curve breakpoints set by the network operator based on the local voltage profile.

Worked example: cos-phi(P) in practice

Suppose an inverter is rated 10 kVA and its network operator specifies a cos-phi(P) characteristic requiring a power factor of 1.0 below 50 percent of rated active power output, linearly tightening to 0.90 lagging (absorbing reactive power) at 100 percent output.

1. At 3 kW output (30 percent of rated), the inverter is below the 50 percent breakpoint, so it operates at unity power factor: reactive power output is 0 kvar.
2. At 8 kW output (80 percent of rated), the inverter is 60 percent of the way from the 50 percent breakpoint to 100 percent, so power factor interpolates from 1.0 toward 0.90: cos-phi = 1.0 minus 0.6 x (1.0 minus 0.90) = 0.94.
3. From cos-phi = 0.94, phi = arccos(0.94) = 20.0 degrees, so tan(phi) = 0.364.
4. Required reactive power Q = P x tan(phi) = 8 kW x 0.364 = 2.91 kvar, absorbed (inductive), pulling local voltage down slightly while the array is producing near its peak.
5. Apparent power S = square root of (P squared + Q squared) = square root of (64 + 8.47) = 8.51 kVA, comfortably inside the inverter's 10 kVA rating, confirming this operating point does not require active power curtailment to stay within the inverter's apparent power limit.

This is the calculation an inverter's firmware performs continuously as active power output changes through the day, and it is also the calculation an engineer should reproduce by hand when validating that a settings sheet actually matches the network operator's specified curve rather than a generic factory default.

Sizing consequences

Because reactive power draws on the same apparent power capacity as active power (S squared = P squared + Q squared), an inverter required to supply significant reactive power at high active power output needs oversized apparent power capacity relative to its DC-side rating, or it must curtail active power slightly to stay within its kVA limit at the moment reactive demand peaks. This is why some network operators specify inverter oversizing (for example, a kVA rating 5 to 10 percent above the array's peak DC-to-AC potential) specifically to preserve room for mandatory reactive power support without forcing active power curtailment.

What goes wrong in practice

The most frequent misconfiguration is leaving an inverter at a factory default fixed power factor of 1.0 (no reactive support at all) when the connection agreement specifies a cos-phi(P) or Q(U) characteristic; this typically passes basic commissioning checks that only confirm the inverter energises and produces power, without verifying it responds correctly to the required curve under varying output. Verifying this requires either logging inverter reactive power output against active power output over a full day, or specifically forcing several test power levels during commissioning.

A second issue is inconsistent voltage measurement location. Q(U) droop control depends on measuring voltage at, or close to, the point of common coupling; some inverters measure voltage only at their own internal terminals, which can differ meaningfully from the connection point voltage due to voltage drop across the site's own internal wiring, especially on longer DC or AC runs within a large commercial installation. This produces a control loop responding to the wrong voltage signal.

A third and often overlooked issue is interaction between multiple inverters on the same feeder all running Q(U) droop independently without coordination; if each responds to the same shared voltage measurement point with an aggressive droop gain, the combined reactive power swing can itself cause voltage oscillation, effectively creating instability from a mechanism intended to improve stability. Network operators addressing high-penetration feeders sometimes specify a shared, slower droop gain, or move to centrally coordinated reactive power dispatch, precisely to avoid this.

Finally, reactive power support is sometimes confused with active power curtailment for the purpose of a connection agreement; a network operator requiring reactive power support is not thereby waiving any separate export limitation requirement covered in the companion article on export limitation and power control, and both mechanisms can be in force simultaneously on the same connection.

Standards and references

EN 50549-1 specifies the reactive power capability and control mode requirements, including the cos-phi(P) and Q(U) characteristic structures, for generators connecting at low voltage across much of Europe. IEC 61727 provided earlier, simpler power factor requirements for smaller PV systems and remains a useful historical reference point. IEEE 1547, in its 2018 revision, introduced comparable voltage regulation categories and reactive power modes for the North American context, aligning the general philosophy with European practice even where specific curve parameters differ.

For the surrounding connection process, see grid connection and export rules explained and the grid application checklist.

Frequently asked questions

Does providing reactive power reduce my exported energy?

Not directly, but it can reduce the inverter's available active power headroom at high output if the inverter is not oversized in apparent power terms, because active and reactive power share the same apparent power limit. A correctly sized inverter, as discussed above, is specified with enough kVA headroom that mandated reactive power support at peak output does not force active power curtailment.

Why does my inverter absorb reactive power instead of exporting it?

Most cos-phi(P) and Q(U) requirements are configured to absorb reactive power (operate inductively) at high output specifically because PV export tends to raise local voltage, and absorbing reactive power counteracts that rise. Exporting reactive power (capacitive behaviour) would push voltage up further, which is rarely what a network operator wants from PV specifically, though it may be requested during low-load, low-voltage periods.

Is reactive power support the same as power factor correction for my own building?

No. Building-side power factor correction addresses the site's own load characteristics for billing or equipment reasons, and is typically implemented with separate capacitor banks or drive-based correction. Inverter reactive power support responds to the network operator's requirement at the point of connection and can, in some configurations, unintentionally interact with an existing power factor correction system if not coordinated.

Can a battery inverter provide reactive power support when not charging or discharging?

Some hybrid and standalone storage inverters can provide reactive power support even at zero active power output, since the reactive capability depends on available apparent power headroom rather than active power flow. This is sometimes offered as a distinct grid-support service and should be explicitly checked on the inverter's datasheet rather than assumed.

How do I know which control mode my network operator requires?

The connection offer or the network operator's published grid code profile document specifies the required mode and, for cos-phi(P) or Q(U), the exact breakpoints and gain. Installers select the matching preset in the inverter firmware, and it should be documented on the commissioning certificate alongside the protection settings described in the companion article on grid codes and inverter protection settings.

Does reactive power support cost the system owner anything?

Directly, no additional equipment cost is usually required beyond correct inverter sizing and configuration, since the capability is built into standard grid-tied inverters. Some jurisdictions offer or require compensation arrangements for reactive power services at larger scale, but for typical residential and small commercial PV, it is treated as a standard condition of connection rather than a separately paid service.