Installations with self-supply and their effect on power factor correction systems

The rapid development of the main components of self-supply photovoltaic systems towards more efficient and less expensive devices (both in terms of the photovoltaic generation itself and the inverters for DC-AC conversion), as well as the reasonable and expected changes to national laws (which have considerably streamlined and simplified the procedures and requirements for photovoltaic self-generation), are resulting in a significant and unstoppable increase in the implementation of self-supply systems. And this is happening not only in residential settings, but with larger energy consumers at the industrial and service level.

However, from the point of view of the power factor correction systems that are or will be installed, it is important to consider a series of factors that may cause these systems to malfunction. This could even result in penalties for excess reactive power consumption that were not being incurred before the self-supply system was installed.

The installation of a photovoltaic power generation system for self supply can cause existing reactive power compensation devices to malfunction. It is important to know these consequences in order to avoid unexpected penalties.

Practically all the problems that may arise are rooted in a basic concept. When installing a self-supply system, we reduce, depending on the power injected by this system at any given time (kW.h), the active energy (kWh) consumed from the grid (the one recorded by the electricity company's meter) compared to what we would consume under the same load conditions without the photovoltaic generation.

Depending on the percentage of autonomous generation with respect to the total consumption of our loads, a number of problems can arise involving reactive power compensation, which can basically be summarized as follows:

• A lower excess kvarL·h that is used to determine the reactive power penalty:
  • Excess of kvarL·h without penalty = 0.33 x kW.h total tariff period.

• Possibility of reading errors in the capacitor bank's power factor regulator caused by:

• • A reading of cos phi by the power factor regulator that is different from the value measured by the billing meter.
  • Excessively low current flow through the current transformer(s) that provide(s) the reading to the regulator.
  • A value of cos phi measured by the regulator that is excessively close to 0.

For a better understanding of this matter, we will consider different connection options below, and discuss the possible associated problems. In all cases, we will assume that the inverter is generating active power (kW) only, meaning it is configured to generate a power factor of 1, as currently required by the Spanish low-voltage code. This means that all the reactive power (generally inductive) consumed by the loads is consumed from the grid.

The main risks associated with this connection mode would be:

• A lower excess kvarL·h that is used to determine the reactive power penalty.
• A reading of cos phi by the power factor regulator that is different from the value measured by the billing meter.

The following example seeks to clear up the reason for these potential problems by considering an installation with the following stable consumption conditions:

• kW III consumed by the loads: 110 kW
• kW III generated by self-supply: 100 kW
• kvarL consumed by loads: 93 kvarL
• Existing capacitor bank: 2 x 10 + 4 x 20 kvarC (80 kvarC)
• cos phi read by the regulator (with 80 kvarC connected): 0.99 L

• cos phi with the self-supply injecting, read by the utility's meter: 0.61

• Excess reactive energy times: 13 kvarL.h - (10 kvarL.h x 0.33) = 9.7 kvarL.h
• Network voltage: 3 x 400 VAC
• Assuming 320 h/month (P1 to P5):

• • Total surcharge in monthly bill: 9.7 kvarL.h x 320 x €0.062332 = €193.5

This connection mode may pose some problems in the operation of the capacitor bank. The current transformers that provide a signal to the automatic regulator only read the active power consumption not generated by the self-supply, and not the entire reactive power of the loads that need to be compensated. If the percentage of active power contributed by the self-supply is low or medium with respect to the total consumed by the loads, the regulator will very likely work normally. If, however, the self-generation accounts for a high percentage of the total consumption, there is a high risk that the regulator will malfunction.

The main risks associated with this connection mode would be:

• A lower excess kvarL·h that is used to determine the reactive power penalty.
• A reading of cos phi by the power factor regulator that is different from the value measured by the billing meter.

• A value of cos phi measured by the regulator that is excessively close to 0

The following example aims to clarify the reason for this possible malfunction of the regulator, assuming once again an installation with the following stable consumption conditions:

• kW III consumed by the loads: 110 kW
• kW III generated by self-supply: 102 kW
• kvarL consumed by loads: 93 kvarL
• Existing capacitor bank: 10 + 4 x 20 kvarC (90 kvarC)
• Current flow through the CT (with 90 kvarC connected): 12 A

• If we assume a CT of 400/5 A, 12 A of current may not provide a suitable measurement signal to the regulator, in terms of either power and/or precision ⇒ The regulator disconnects the steps due to detecting a critical alarm situation due to the absence of a current reading.
• With the steps disconnected, the cos phi read by the regulator is: 0.08 L
• High possibility of unstable regulator operation, which means that the reactive power compensation is ineffective, with the risk of incurring penalties due to excess consumption of inductive reactive power that did not exist before.

One possible solution is to at least try to avoid regulator malfunctions, and relies on converting type #2 to type #1 to ensure that the cos phi measured by the regulator is the actual cos phi of the installation based on the total kW consumed by the loads needing compensation, regardless of whether it is consumed from the grid or is self-generated.

To do this, one option is to install other CTs and measure the power generated by the self-supply and take their secondary signal, together with the CTs at the input, to a 5+5/5 A current adding transformer (CIRCUTOR model TSR-2), whose secondary signal will provide the current measurement to the power factor regulator. As shown in Connection Type #3.

Advanced compensation method

The most effective solution, which would ensure proper reactive power compensation regardless of the operating status, characteristics and connection point of the self-supply system, is to install SVGm static VAR generators, as a complementary compensation system.

A detailed analysis of the compensation needs will help us determine the most suitable SVGm device to include in our installation and avoid any penalties that may arise and that are not avoidable with a conventional capacitor bank system. It will also help us determine the economic feasibility of its installation.

The main features of these SVGm static VAR generators are as follows:

• Ability to inject reactive power (compensation) in any current condition through the CTs.
• Precise adjustment of the reactive power injected to achieve the target cos phi, since this system does not rely on capacitor banks.
• Immediate response to load variations, free of transients, and immune to the presence of harmonics in the network.

One possible way to connect the SVGm device in our electrical network is that indicated in Type # 4, but there are others, depending on the unique requirements of each installation, which must be analysed specifically for each case.

General conclusions on reactive power compensation in installations that feature a self-supply system

We could list the following main considerations on this issue:

• The reduction in active energy (kWh) counted by the billing meter is the source of the main problems that may occur; therefore, the higher the percentage of generation from self-supply compared to the total consumption of the installation (receivers), the greater the likelihood of these problems occurring.
• Depending on the cause and seriousness of the problem with reactive power compensation, different solutions can be considered. These may include:

• • Changing the injection point for the self-supply to the network.
  • The addition of current transformers to measure the power injected to the network.
  • Changing the power of the existing capacitor bank:
    • Increasing the capacity bank power to achieve an average total cos phi as close as possible to 1 in different load consumption and self-supply injection conditions.

• Adding smaller power steps than the smallest current step in the bank to allow it to compensate for lower kvarL consumption, which now could entail a penalty. Basically, have smaller power steps for a more precise adjustment
• Installing individual compensation in machinery or in sub-panels to supplement the general reactive power compensation provided by the automatic capacitor bank at the main connection.
• Outfitting the bank with a regulator from the Computer SMART III range that is able to measure the reactive energy consumed by the 3 phases, just like the utility meter does, to achieve a more precise reactive power compensation.
• Use, in any case, a regulator that can measure energy in all 4 quadrants, that is, both when the installation is consuming and producing electricity. This feature is especially important in the case of self-supply systems that can export electricity to the grid. Although it only appears that consumption is registered in the consumption quadrants (Q1 & Q4), we wholeheartedly recommend their use both to anticipate possible future changes in the invoicing system, and to avoid the faulty operation of the regulator. In this regard, remember that all CIRCUTOR regulators in the most recent ranges include measurement in all 4 quadrants.