A 60 MW solar plant at the end of a 40 km radial line fails commissioning for one reason: voltage. At noon the inverters push full output into a high-impedance feeder, the point of connection climbs past the upper statutory band, and the utility raises the alarm before the plant earns a day of tariff.
The owner installs capacitor banks, the standard first attempt. They correct the daytime condition, over-correct after sunset, and trip on overvoltage at night. Step switching cannot follow a source that moves 30 % of its output in seconds.
An MV STATCOM for PV plants solves the problem at the point of connection instead of downstream of it. It holds voltage inside the band continuously, absorbs as well as injects reactive power, and reaches full output in one cycle. This article covers the physics, the grid code duties, and the selection logic.
Voltage at the point of connection follows ΔV ≈ (P·R + Q·X)/V. On a long radial feeder the X term dominates, and a low short-circuit ratio makes the bus soft enough to move with every load step.
The same impedance is now carrying power in the reverse direction. A feeder designed to deliver power outward carries it inward at unity power factor, which produces the highest voltage the plant will ever see — at its own maximum output.
Cloud transients add a second duty. A passing cloud moves 20–30 % of plant output within seconds, and the bus voltage swings with it. Inverters trip on overvoltage, generation is lost, and the operator records a disturbance the plant caused.
Voltage regulation is the visible requirement, not the only one. Connection studies commonly specify a reactive range equivalent to ±0.95 power factor at rated output, continuous from full capacitive to full inductive, with no step behaviour.
Low-voltage ride-through adds a dynamic duty. During a voltage dip the plant must inject reactive current in proportion to the depth of the dip, typically within one to two cycles of detection. A switched device is still counting contactor operations at that point.
Harmonic and flicker limits bound the design from the other side. IEEE 519-2014 limits current distortion, and THDi at the point of connection is reported alongside Total Harmonic Distortion measurements taken during commissioning. Flicker limits are verified against Pst values recorded during cloud transients.
A capacitor bank cannot satisfy those three duties at the same time, because two of them are dynamic and one of them changes sign.
Both devices use a voltage-source converter and both compensate reactive power continuously. The difference is where each one connects and how much current it can move.
A low-voltage SVG sits behind a step-up transformer on the LV bus. An MV STATCOM for PV plants connects directly to the 11 kV, 33 kV or 36 kV bus that the utility actually regulates, so the reactive current travels the shortest possible path to the point of connection.
|
Parameter |
LV SVG (400–690 V) |
MV STATCOM (11–36 kV) |
|
Connection to the point of connection |
Through step-up transformer and LV switchgear |
Direct on the MV bus, no step-up stage |
|
Typical unit rating |
30 kVar to roughly 1 MVAr |
A few MVAr up to 100 MVAr and above |
|
Output current during a voltage dip |
Current-source behaviour, holds output at low voltage |
Same behaviour, holds current down to roughly 0.3 pu |
|
Where voltage is regulated |
Indirectly, at the LV side of the transformer |
Directly, at the bus the utility measures |
|
Conversion losses |
Transformer, cable and converter losses |
Converter and cooling losses only |
|
Cooling |
Forced air |
Forced air at low rating, water cooling above a few MVAr |
|
Footprint |
Wall-mounted cabinet or panel |
Container or dedicated electrical building |
|
Typical application |
C&I rooftops, LV distribution, small solar |
Utility PV, wind, transmission-connected plants, weak grids |
Table A — LV SVG versus MV STATCOM (typical figures; confirm against the project specification).
Low-voltage SVG remains the correct answer for a factory bus. The choice is not technology preference; it is the voltage level at which the utility measures and the current the plant must move to change it.
Output quality is set by the inverter topology: a three-level NPC stage divides the DC link, so each switching step is Vdc/2 rather than Vdc. That halves the dv/dt seen by the machine and by adjacent control wiring, and it lowers the PWM harmonic content at the same switching frequency, which shrinks the output filter. SPWM modulation with phase-shifted carriers spreads the switching spectrum further.
IGBT modules still dominate this power class because the current is high and the switching frequency is moderate. SiC MOSFET devices are raising the switching-frequency ceiling in lower-current designs, but they do not yet displace IGBT modules at medium voltage. What separates one STATCOM from another is the thermal design behind those modules, and thermal stress, not switching stress, is what ends the life of a power stack.
A container sitting in a Gulf or Indian summer sees 45 °C ambient and solar gain on the enclosure roof. A design rated for full output at 45 °C ambient, with water cooling above a few MVAr, delivers its nameplate current in July. A design that derates above 40 °C does not, and the shortfall appears precisely when the plant needs reactive support most.
Size from the reactive power the grid operator demands at the point of connection, not from the plant's DC capacity. The connection study states the range; the short-circuit ratio tells you how hard the STATCOM will work to hold it.
Short-circuit ratio is the fault level divided by the plant rating. Above roughly 10 the bus is stiff, and a modest dynamic range satisfies the reactive requirement. Between 3 and 5 the bus is weak, voltage rise becomes the binding constraint, and the unit must absorb reactive power at full solar output — the exact case a capacitor bank makes worse.
Confirm the overload requirement before you finalise the rating. A short-time overload capability, commonly specified around 1.1 pu for one hour, lets a smaller STATCOM ride through transient conditions instead of being sized for the worst second of the year.
|
Plant or bus condition |
Short-circuit ratio at the point of connection |
Indicative dynamic range |
Cooling |
Notes |
|
50 MW PV, short feeder, PF compliance only |
Above 10 |
10–15 MVAr |
Forced air |
Voltage regulation is the only duty |
|
100 MW PV, 40 km radial line |
3–5 |
25–35 MVAr |
Water |
Voltage rise at noon plus dip support |
|
PV with strict flicker limits |
5–10 |
20–30 % of plant rating |
Water |
Dynamic bandwidth is the driver |
|
Wind or hybrid plant on a weak bus |
Below 5 |
25–40 % of plant rating |
Water |
Ride-through reactive current injection |
|
Industrial MV bus with arc furnace or rolling mill |
Not applicable |
10–30 MVAr |
Water |
Flicker, unbalance and voltage dips |
Table B — Indicative sizing matrix (figures are indicative; confirm against the connection study and project specification).
A modern MV STATCOM controls more than the fundamental. Digital extraction of the load spectrum lets it compensate selected harmonic orders from the 2nd to the 50th, plus negative-sequence current from unbalance.
Treat that as a secondary function, not a substitute for a dedicated active harmonic filter. If measured harmonic current is large relative to the STATCOM rating, the harmonic duty consumes capacity the voltage regulator needs.
Install the current transformers so they measure the plant without the STATCOM's own output. CT polarity errors are the most common cause of a compensator that appears to make voltage regulation worse.
Delivery is containerised, which moves most of the work into the factory. The site scope is the MV connection, the cooling circuit, the CT wiring and the control interface to the plant SCADA.
Commissioning is a measurement exercise. The utility typically witnesses response time, the reactive range across the full output curve, and ride-through behaviour, and it may ask for a PSCAD or PSS/E model before the connection is approved.
Maintenance is limited to cooling components. Fans, air filters and water quality are the wear items, inspected annually in clean rooms and quarterly in dusty sites. There are no contactors to weld and no capacitor cans to replace, which is the operational argument that outlives the capital cost.
An MV STATCOM for PV plants is the correct answer when the plant connects to a weak bus, when the utility regulates voltage at medium voltage, and when the reactive duty changes sign across the day. It holds the point of connection inside the statutory band, absorbs reactive power at full output, injects reactive current during a dip, and does all three without a step.
Size it from the connection study, confirm the ambient and the overload requirement, and match the cooling to the site. Those three decisions determine whether the plant meets its grid code obligations in the first year and in the tenth.
1. How do I size a STATCOM for a solar plant?
Start from the reactive range in the connection study, not from the plant's DC capacity. Then check the short-circuit ratio at the point of connection: below 5 the voltage-rise constraint usually drives the rating upward. Add the specified short-time overload capability and confirm the figure with a load-flow study before ordering.
2. Does installation require a plant shutdown?
The unit connects in parallel and never sits in the generation path. Most of the work happens in the factory, so site time is limited to the MV connection, cooling circuit, CT installation and control wiring. That normally fits a scheduled maintenance window rather than a generation stop.
3. Should I choose water cooling or forced air?
Choose by rating and climate. Forced air is sufficient at low ratings in temperate conditions. Above a few MVAr, or in a container exposed to 45 °C ambient, water cooling keeps the power stack at full output instead of derating through the hottest weeks of the year.
4. How does a STATCOM differ from an SVC on the same bus?
A STATCOM is a voltage-source converter, so its output current is largely independent of system voltage and its reactive range is symmetrical, capacitive to inductive, at full speed. An SVC relies on thyristor-switched impedances, and its reactive output falls with the square of the voltage — exactly when the grid needs support most.
Subscribe to us to enjoy event prices and get some of the best prices.
IPv6 network supported