Leading-edge technology and professional expertise

FRAKO offers you complete systems from a single source to optimize your power network and improve your power quality with active and passive harmonic filters, dynamic power factor correction systems and network analysers.

Power factor correction for solar parks

Author: Peter Riese

In recent decades, the drive to reduce our dependence on fossil fuels for power generation has resulted in solar parks, otherwise known as photovoltaic power stations or simply PV systems, establishing themselves as a reliable source of electrical energy. In the case of the larger systems, maximization of their overall efficiency is an important aspect in ensuring that the power generated remains not only reliable but also affordable. In addition to the usual measures taken to enhance the performance of the solar modules themselves, there is further potential in optimizing the energy they produce.

In Germany, most solar parks typically generate between 1 MW and 20 MW, and are thus governed by the technical regulations of the Federal Association of the German Energy and Water Industries (BDEW): ‘Directive for the connection and parallel operation of generation plants in the medium voltage network’.

Since 2008, this directive has set out the key points to be observed when a power generation system is connected up to a power distribution operator's medium voltage network.

These requirements include a network-compatible power factor at the feed-in point. This means that when delivering active power, the generation plant must be capable, under all load conditions, of operating with a reactive power corresponding to a power factor cos phi between 0.95 lagging and 0.95 leading at the network entry point. Overexcited operation (capacitive) would increase the voltage, while underexcited operation (inductive) would reduce it; generation systems must therefore help in maintaining a steady network voltage.

Various network configurations and load conditions give rise to a variety of needs, which is why power distribution operators stipulate individual requirements for the network section concerned. These can vary greatly from one operator to the next, and are set out in the form of a characteristic curve where the power factor is stated either as a function of the active power (the cos phi (P) characteristic) or directly as a function of the voltage (the Q(V) characteristic). The BDEW directive calls for the power factor to be regulated to the target value thus stipulated within 10 seconds.

Optimization potential 1:

Modern inverters are ‘reactive power capable’ and can convert the energy captured by solar modules and feed it into the utility network in compliance with the characteristic curves. However, there is a great disadvantage to this: inverters deliver apparent power, i.e. the geometric sum of active and reactive power! The worse the required power factor specified, the less is the revenue-earning active power that an inverter can supply, regardless of whether this in the inductive or capacitive region.

Example: If a power factor of 0.95 is stipulated, the inverters must supply an amount of reactive power equivalent to 33% of the active power. However, if this required reactive power is supplied by a suitable power factor correction system, allowing the inverters to operate at a power factor of 1, this yields 5% more billable active power for the same inverter apparent power!

To make it at all possible to parallel larger PV systems to existing medium voltage networks, some present-day operators are even specifying power factors that can go as low as 0.90. In this case, the amount of reactive power required is equivalent to 48% of the active power, and the billable active power exported can thus be increased by 11%!

It is therefore a worthwhile proposition to compensate the reactive power separately, so that the inverters can deliver the maximum possible active power!

Optimization potential 2:

When modern inverters operate under no-load conditions (particularly at night or if the solar panels are deprived of sunlight by clouds, fog, snow, etc.), the total no-load reactive power from all the inverters puts a heavy capacitive load on the network.

To suppress this effect, inverters are often switched off during such periods, but this considerably reduces their service life. In this case as well, a separate power factor correction system offers significant potential for system optimization.

Optimization potential 3:

The length of the cabling between a solar park and the network entry point can easily amount to several kilometres. As the conductors are laid close together, long underground cables have a capacitive effect, whereas overhead lines have an inductive effect. One kilometre of underground cabling can therefore give rise to several kvar of capacitive reactive power, which, since it develops on the way to the solar park and thus can neither be detected nor measured by the inverter control system, must be corrected separately.

The problem:

The network operator's specifications relate to the PV system's entry point to the medium voltage network, which is frequently several kilometres away from the solar park itself. This means that these specifications and the BDEW directive must be complied with not directly at the solar park, but at the feed-in connection!

The solution:

Some time ago, FRAKO carried out an in-depth study of a medium-sized solar park, applying the three optimization approaches outlined above. An appropriate power factor correction system was then designed, installed and commissioned.

Solar park key data:
- System rating: 3122 kWp
- Annual output: 2 849 400 kWh
- Surface area: 4.7 hectares
- AnnualCO2savings:1710tonnes
- Distance of park from feed-in point: 5 km

Network operator's specifications:

- Target power factor 0.95 inductive during network feed-in operation

- The power factor must never be capacitive when power is drawn from the network.

- Regulation of the power factor to the target value within 10 seconds as per the BDEW directive.

These specifications lead to the requirement that appropriate power factor correction is made—during the day for the necessary dynamic reactive power of the entire solar park plus the cabling capacitance, and at night for the no-load reactive power of the inverters plus the cabling capacitance.

Reactive power determined:

1) The reactive power needed to correct cos phi at the inverters to 1.00 amounts to a total of 950 kvar inductive (equivalent to about 1/3 of the active power).

2) Total no-load reactive power of the 141 inverters: 98 kvar capacitive

3) Cabling reactive power under partial load and no load conditions: 45–90 kvar capacitive.

As all three of the above circumstances call for inductive reactive power to be provided, a power factor correction system with a total of 990 kvar inductive reactive power was installed, distributed across seven bayed cabinets. The system was designed with differently sized stages, comprising low-loss inductors, switchgear and group overcurrent protection.

It was essential to measure and evaluate the various requirements for reactive power. This was done by linking measurements at the remote feed-in point with those directly in the solar park and transmitting their readings to an intelligent reactive power control relay, which switched the stages of the power factor correction system in or out within the stipulated 10 seconds, thus ensuring ongoing compliance with the specifications.


Thanks to the additional yield of billable active power achieved by the power factor correction system, the investment paid for itself completely within 22 months, and now continues to boost the revenue of the solar park by 5%.

November 2018