Synchronous Condensers vs. Capacitor Banks for PF Correction
Both synchronous condensers and capacitor banks inject reactive power into an AC system, but the similarity stops there. One is a spinning machine that responds dynamically to grid conditions; the other is passive hardware that delivers a fixed block of VARs the moment you close a breaker. Choosing between them comes down to the scale of the problem, the speed of load swings, and how much capital you can justify.
How Each Technology Supplies Reactive Power
A capacitor bank stores energy in an electric field and releases it 90 degrees out of phase with the supply voltage, producing leading reactive current. The output is proportional to voltage squared, so a bank rated at 1 MVAR at 11 kV delivers only about 0.82 MVAR if the bus drops to 10 kV. That voltage dependence is a real limitation during fault conditions, exactly when reactive support is needed most.
A synchronous condenser is a synchronous motor running at no mechanical load. Its field winding is excited by a DC exciter, and by adjusting excitation current the operator controls whether the machine absorbs or supplies reactive power. Over-excite it and it pushes VARs into the network; under-excite it and it draws them. Output tracks a command signal continuously, independent of bus voltage (within rated limits). The machine's rotating inertia also adds short-circuit current, which helps protection systems operate reliably.
For a deeper look at what power factor actually measures and why it matters, the relationship between real, reactive, and apparent power is worth understanding before selecting correction hardware.
Fixed vs. Stepped Capacitor Banks
Capacitor banks come in three configurations. Fixed banks are the simplest: a single switched block, always on or always off. They suit steady, predictable loads like induction motors running at constant speed. Stepped banks add a contactor or thyristor-switched stages, typically 50-100 kVAR per step, so the correction tracks load changes in discrete increments. Automatic power factor correction (APFC) controllers read power factor on the bus and switch stages in or out, usually targeting 0.95 or better.
The limitation of any capacitor bank is that correction happens in steps, not continuously. A load that swings quickly between light and full duty can temporarily overcorrect between switching events, causing a leading power factor that creates its own problems. You can read more about those risks in the article on overcorrection and leading power factor.
Automatic power factor correction panels are the practical implementation of stepped banks for industrial sites; they handle the switching logic and protect against resonance conditions that plain fixed banks ignore.
Synchronous Condensers: Continuous and Bidirectional
The key operational advantage of a synchronous condenser is its continuously variable, bidirectional output. A 20 MVAR machine can deliver anything from -20 MVAR (absorbing reactive power, useful during light-load conditions when capacitance already on the network threatens voltage rise) through zero to +20 MVAR. That range, controlled by the automatic voltage regulator (AVR), is why grid operators and large industrial sites with arc furnaces or rolling mills prefer synchronous condensers for voltage regulation.
Response speed differs sharply from capacitor banks. A thyristor-switched capacitor stage responds in roughly one cycle (16-20 ms at 50-60 Hz). A synchronous condenser with a modern AVR responds within a similar window for small changes but takes longer for large excursion demands because the exciter field time constant is in the range of 0.1-0.5 seconds. For truly fast transients, static VAR compensators (SVCs) or STATCOMs are faster, but synchronous condensers beat capacitor banks on dynamic performance.
Cost, Maintenance, and Losses
Capacitor banks are inexpensive per MVAR. A 500 kVAR fixed bank for a distribution panel might cost $3,000-$8,000 installed. A stepped 2 MVAR APFC panel with 20 stages runs $40,000-$80,000. There are essentially no moving parts, maintenance is limited to periodic inspection and capacitor replacement (typical life 15-20 years), and no-load losses are very low, usually under 0.5% of rated kVAR.
Synchronous condensers carry a substantial capital premium. A 50 MVAR machine installed at a transmission substation might cost $10-25 million including foundations, cooling, and protection systems. Maintenance intervals mirror those of large motors: bearing inspections, brush/slip-ring service, periodic stator winding tests. No-load losses from friction, windage, and core losses typically run 1-3% of rated capacity, a continuous operating cost that adds up over decades.
That said, the cost comparison is not simply "capacitors are cheaper." A very large installation requiring dynamic response and bidirectional capability might need 10+ banks of thyristor-switched capacitors plus a STATCOM to approximate what one synchronous condenser does, and the total cost can converge.
Harmonic Behavior
Capacitor banks interact with system inductance to create resonant circuits. At certain frequencies, the impedance of a bank-plus-transformer combination drops sharply, amplifying harmonic currents that variable-frequency drives, arc furnaces, and rectifiers produce. This is a genuine design concern; sizing power factor correction capacitors correctly requires accounting for harmonic spectrum, not just fundamental-frequency reactive demand. Detuned reactors (typically tuned to the 4.7th or 7th harmonic) are standard additions to capacitor banks on sites with significant nonlinear loads.
Synchronous condensers have a natural advantage here. The machine's synchronous reactance presents a low impedance path for harmonics, damping them rather than amplifying them. This is one reason they remain attractive at the transmission level, where large converters and HVDC links generate substantial harmonic content.
Comparison at a Glance
| Feature | Capacitor Bank | Synchronous Condenser |
|---|---|---|
| Reactive output | Fixed or stepped (discrete) | Continuously variable |
| Response speed | 1 cycle (thyristor) to seconds (contactor) | ~100-500 ms (large step) |
| Bidirectional (absorb/supply VARs) | Supply only | Both |
| Voltage-dependent output | Yes (output falls with voltage) | No |
| Harmonic resonance risk | High without detuning | Low (acts as damper) |
| Capital cost (per MVAR) | Low-moderate | High |
| Maintenance | Low | Moderate-high |
| Short-circuit contribution | None | Significant |
| Typical application size | kVAR to tens of MVAR | 20 MVAR to 400+ MVAR |
Where Each Fits
For most industrial plants, capacitor banks are the right answer. A plastics injection-molding facility running constant-speed motors at 0.78 power factor can correct to 0.95 with a fixed or lightly stepped bank sized from the standard correction formulas. Cost is low, payback on reduced utility demand charges is fast (often 1-3 years), and the load profile is stable enough that fixed correction works.
Arc furnace shops, steel mills with large rolling lines, and facilities with frequent motor starts are better candidates for dynamic reactive compensation, but even here a hybrid approach often wins: a base capacitor bank handles the average reactive load cheaply, and a smaller dynamic device (synchronous condenser or STATCOM) handles the swing component.
At the grid and transmission level, synchronous condensers have seen a resurgence. As thermal generators retire, they take their rotational inertia and short-circuit current with them, leaving networks that struggle during faults. Grid operators in the UK, Australia, and the US have commissioned new synchronous condensers specifically to replace the inertia and fault current that solar and wind cannot provide. In that application, the reactive power compensation is almost secondary to the stability services the machine delivers.
Frequently Asked Questions
Can a synchronous condenser replace a capacitor bank at an industrial plant?
Technically yes, but economically it almost never makes sense below about 10-20 MVAR of requirement. The capital and maintenance costs of a synchronous machine far exceed those of a capacitor bank for the steady correction needs of typical industrial loads. Reserve synchronous condensers for applications demanding continuous control, bidirectional operation, or fault current contribution.
Do capacitor banks cause harmonic problems on every site?
Not automatically. Sites with mostly linear loads (resistive heating, constant-speed motors) rarely see harmful resonance. Problems arise when the site has significant harmonic-generating equipment and the capacitor bank happens to resonate near a dominant harmonic order. An engineer should check the impedance frequency scan before specifying any bank on a site with drives, rectifiers, or arc equipment.
How long does a synchronous condenser last compared to a capacitor bank?
Well-maintained synchronous condensers routinely operate for 40-50 years; the machines installed in the 1970s at many substations are still running. Individual capacitor units have a shorter life (15-20 years is typical), though staged replacement means a bank can run indefinitely with periodic cell swaps. On a lifecycle cost basis the gap between the two technologies narrows considerably over long time horizons.
What is a "retrofit" synchronous condenser?
Grid operators sometimes convert retired synchronous generators into synchronous condensers by decoupling the prime mover (removing or disconnecting the turbine/steam path) and adding the excitation controls needed for VAR regulation. This is cheaper than a new build and reuses existing substation infrastructure. The practice has become common in Europe and Australia as part of grid stability programs.