SVG vs Capacitor Bank: Which Reactive Compensation Fits?

Summary. Choosing between an svg vs capacitor bank decision is really a question about how stable the load and the network are. A capacitor bank is a fixed, inexpensive source of reactive current that performs well on steady loads with a stiff supply; a static var generator is a converter that produces continuously variable reactive current and holds performance when the load fluctuates, the voltage sags or the harmonics are high. This guide compares the two technologies on response, harmonic interaction, resonance risk, footprint, losses and lifetime cost, explains where a hybrid arrangement using both is the sensible engineering answer, and sets out the measurements that settle the question for a specific plant.

What Is the Technical Difference Between an SVG and a Capacitor Bank?

A capacitor bank is a passive device. Its reactive output rises with the square of the applied voltage and is fixed by the installed kVAr, so the compensation it provides depends on how many stages the controller has switched in and on the voltage at that moment. It is simple, robust, inexpensive and, on steady loads with a stiff supply, entirely adequate.

A static var generator (SVG) is a voltage-source converter that synthesises reactive current continuously. It measures voltage and current, calculates the reactive component and injects leading or lagging current within its rating, with no discrete steps. Output is essentially independent of voltage across the operating range, which is why an SVG continues to support the network during a dip that would collapse a capacitor bank's output.

The difference in principle produces differences in practice. A capacitor bank with a detuned reactor is tuned to a specific frequency and protects itself against resonance, but it cannot avoid interacting with the network: the combination of capacitance and source impedance always creates a resonant point, and if harmonic current appears near that point the bank amplifies it. An SVG is a current source, so it has no resonant point of its own and does not amplify harmonics.

Speed is the second practical difference. Contactor-switched stages respond in tens of milliseconds at best, and each switching event produces a step in reactive output that can be visible as a voltage step. A converter responds within a few milliseconds and produces continuous output, which is why flicker suppression on welding, hoists and arc loads is normally an SVG application rather than a capacitor application.

Losses and lifetime differ as well. A capacitor bank has very low losses when healthy but degrades with temperature, harmonics and overvoltage, and its capacitance falls over time. An SVG has converter losses that are constant with output but a steady state that can be monitored, and its wear items are the DC bus capacitors and cooling fans. Neither is universally better; each fails differently and at a different point in its life.

What "detuned" actually means. A detuned reactor is connected in series with the capacitor so that the branch presents an inductive characteristic at the dominant harmonic, typically tuned below the 5th order. This protects the capacitor from harmonic current but does not make the installation harmonic-free, and it reduces the effective reactive output because the reactor consumes reactive power. A bank quoted as 100 kvar with reactors may deliver notably less than 100 kvar at fundamental frequency, which buyers should confirm before comparing with a converter rating.

Why the Choice Changes Economics and Performance

1. Response decides whether flicker is solved. Flicker limits are assessed over short intervals. A bank that switches in discrete steps cannot follow a fluctuating load, whereas an SVG compensates continuously and reduces the measured flicker severity. On sites with welding, hoists or arc processes, the bank is simply the wrong tool.

2. Harmonic interaction is a real risk with capacitors. Adding capacitance changes the network resonance. If the plant already has significant harmonic current, the bank must be detuned, and even then the tuning is valid only for the assumed spectrum. A low-voltage SVG for dynamic PF correction and harmonic suppression avoids the resonance question altogether.

3. Compensating a fluctuating load without over-compensating. A bank controlled by a power factor relay tends to hunt: it switches up, overshoots, switches back. An SVG with a defined target produces no steps and no hunting, which also removes the voltage steps that sensitive equipment notices.

4. Space and installation differ more than price lists suggest. A bank needs reactors, contactors, fuses and cooling in a dedicated enclosure, plus cable and switchgear, while a compact 15 kvar wall-mounted SVG can serve a commercial board with a fraction of the footprint.

5. Losses matter at high hours of operation. Converter losses are constant, so on a lightly loaded bank installation with long operating hours the passive solution may use less energy — but if the bank is oversized and switched in and out, its contactors and reactors also contribute losses and maintenance.

6. Lifetime cost depends on failure mode. Capacitors fail gradually and can fail catastrophically under resonance; converters fail as modules that can be replaced. For a critical process, replaceable modules with spares on the shelf are often worth more than the lower purchase price of the passive alternative.

7. Hybrid designs capture both benefits. Many plants install a detuned bank for the steady base reactive demand and a converter for the fluctuating component. A modular SVG for real-time voltage stabilisation on distorted grids pairs naturally with an existing bank, provided the resonance check is repeated.

Measuring the comparison fairly. Compare devices at the same point of connection, the same target power factor and the same load profile. A converter quoted in kVAr and a bank quoted in kVAr can deliver different effective reactive current depending on voltage, detuning and staging. Convert both to reactive current at minimum, average and maximum load before concluding that one is cheaper per unit of compensation.

How to Decide for a Specific Plant

Step 1 — Measure the reactive profile. Log power factor over at least a week, at the resolution of the fastest load changes on site. If PF varies within a narrow band, a bank is adequate; if it swings repeatedly through the day, a converter is indicated.

Step 2 — Measure harmonic distortion at the same time. High THD changes the decision because of resonance risk. Where THD is significant, either detune the bank correctly or choose a converter-based solution.

Step 3 — Check the utility's requirement. Some connection agreements require a defined response time or voltage support capability. Where the requirement mentions dynamic support or flicker, the bank cannot satisfy it regardless of size.

Step 4 — Calculate the reactive range, not a single value. Record the minimum and maximum reactive demand. An SVG sized to the range covers the whole profile; a bank covers it in steps and must have enough stages to avoid hunting.

Step 5 — Compare total installed cost. Include switchgear, reactors, cabling, enclosure space and commissioning. A rack-mount SVG from 10 to 150 kvar may be cheaper installed than a bank plus its detuning reactors, particularly in retrofit situations where space is expensive.

Step 6 — Consider staged investment. Where the load will grow, a modular converter can be expanded by adding units. A bank can also be extended, but each extension requires new stages, new protection and a fresh resonance check.

Step 7 — Review the existing installation before replacing it. Where a bank is already installed and healthy, adding a converter for the fluctuating component is often cheaper than removing the bank entirely. A power factor correction unit rated 75–125 kvar can continue serving the base demand while the converter handles dynamics.

Two practical cautions. First, a converter does not remove the need to understand harmonics: if distortion is high, the right answer may be an SVG combined with harmonic filtering rather than a larger converter. Second, verify the flicker or PF target with a measurement after installation, because the improvement depends on where the measurement point sits relative to the compensation.

Practical installation differences. A bank needs space for capacitors, reactors, contactors and fuses, plus ventilation and a discharge path for stored energy. A converter needs space for the same converter family used in harmonic filters, which is often a standard cabinet or rack. In retrofit projects where the electrical room is full, the converter's smaller footprint frequently decides the procurement before the price is discussed.

Protection and safety differences. A capacitor bank stores energy and must include discharge resistors and a means of isolating individual stages for maintenance; contactor or thyristor switching components wear and need periodic inspection. A converter includes its own protection, typically reports faults to a monitoring system, and isolates internally. For plants with limited maintenance resources, the converter's self-diagnosis is often worth more than the bank's lower component count.

A note on sizing discipline. Whatever technology is chosen, size from the measured reactive profile rather than from a target power factor alone. A bank sized to reach 0.95 at peak load may be badly matched overnight, and a converter sized to peak requirement may run at low output for most of the day. The profile, not the headline figure, determines both the rating and the payback.

When the decision is genuinely close. If the reactive profile is flat, distortion is low, space is available and capital is tight, the bank wins on economics and there is no engineering reason to insist otherwise. If any one of those conditions fails — a fluctuating load, high THD, no space or a utility requirement for dynamic support — the converter wins on function. Recognising which condition applies is what separates a considered specification from a technology preference.

FAQ

Is an SVG always better than a capacitor bank?

No. On a steady load with low distortion and a stiff supply, a detuned capacitor bank provides reactive compensation at lower capital cost and with very low losses. An SVG earns its cost when the load fluctuates, the voltage is weak or harmonics are present.

Can an SVG replace an existing capacitor bank?

It can, but it is often not the cheapest route. If the bank is healthy and correctly detuned, keeping it for base reactive demand and adding a converter for the dynamic component usually costs less than a full replacement.

Does an SVG fix harmonic distortion?

Only if it has harmonic compensation capability and sufficient rating. Its primary function is reactive current. Where distortion is the dominant problem, a dedicated active harmonic filter uses converter capacity more efficiently than an SVG in voltage-control mode.

How fast does an SVG respond?

Typically within a few milliseconds, which is fast enough to affect flicker measurements. A capacitor bank responds in tens of milliseconds at best and in discrete steps, which is why it cannot suppress flicker.

What maintenance does an SVG need?

Cooling fan inspection, filter cleaning and periodic checks of bus capacitor condition. Modules are replaceable, so a failure is usually a swap rather than a rebuild. A capacitor bank needs periodic capacitance measurement and contactor inspection.

Does an SVG work during a voltage dip?

It maintains reactive output within its rating down to a defined voltage, which is one of its advantages over a bank whose output falls with the square of voltage. Check the specified operating range for the model under consideration.

Conclusion

The svg vs capacitor bank question is answered by measurement, not by preference. If reactive demand is steady and distortion is low, a detuned bank is the economical choice. If demand fluctuates, voltage is weak, harmonics are present or the utility requires dynamic support, a converter delivers performance the bank cannot. Many plants end up with both, and the resonance check remains mandatory either way.

Not sure which route fits your site? Send us a week of PF and THD data plus the utility's requirement, and we will compare a bank, a converter and a hybrid option on installed cost and performance. Request a reactive compensation review.