Capacitive Switching

Devices for Limiting Capacitor-Switching Transients

The devices used to tame shunt-capacitor switching transients — controlled (point-on-wave) switching, pre-insertion resistors, current-limiting reactors, solid-state switches and the hybrid diode switch. How each limits the inrush, overvoltage or restrike, what it costs, and where it fits, with the timing, sizing and rating figures behind the choice. Follows CIGRE TB 817, with APS engineering interpretation.

Reading time ≈ 15 min

Section 1

Why Capacitor Switching Needs Mitigation

Shunt capacitor banks are switched repeatedly, and hard, for voltage and reactive-power control. Each switching operation can produce a high-frequency inrush current, voltage transients, travelling-wave overvoltages and restrike-related stress, so mitigation is used to reduce the duty imposed on the circuit breaker, the capacitor bank and the connected network. A CIGRE survey found that roughly three-quarters of installed banks exist for voltage support and reactive power, and about 60% of operators switch them once or twice a day — the rest at least weekly. Left unmitigated, each energisation draws an inrush of up to about 5 per unit at kilohertz frequencies, injecting distorted waveforms, voltage transients and travelling-wave overvoltages into both local and remote stations. A capacitor breaker is therefore among the most heavily stressed devices in the network, mechanically and electrically.

This final page brings the series together by comparing the five device families used to reduce that inrush, outrush, overvoltage and restrike-related stress — controlled (point-on-wave) switching, pre-insertion resistors, current-limiting reactors, solid-state switches and the hybrid diode switch — following CIGRE TB 817. No mitigation method is universal: the right choice depends on whether the dominant concern is inrush current, outrush current, remote overvoltage, restrike risk, breaker wear, cost, reliability or continuous controllability. A useful way to compare the families is to ask three questions — what transient each one limits, when it is in the circuit, and what penalty it adds in cost, losses, complexity or availability. For the underlying inrush and recovery-voltage physics see the APS notes on energising capacitor banks and capacitive current switching for HV circuit breakers.

Section 2

Controlled (Point-on-Wave) Switching

Controlled switching reduces the voltage difference across the breaker contacts at the instant of closing or opening: by choosing a favourable point on the voltage wave it lowers the energy injected into the capacitor bank, and so reduces the inrush current and the overvoltage. In practice it delays the making or breaking command so that each pole makes or parts at the optimum instant on the wave, rather than at random. It is the most widely used mitigation for high-voltage capacitor banks — point-on-wave closing is applied to about 60% of HV breakers — and because current and voltage are 120° apart between phases, it works only with single-pole-operated breakers; a three-pole gang-operated breaker adds little unless the poles are mechanically staggered.

Table 1 — What controlled switching buys.

For the	Benefit
Circuit breaker	Lower inrush and electrical wear — less contact erosion and nozzle ablation, longer maintenance intervals and life; controlled opening at a chosen arcing time also lowers the restrike probability by guaranteeing a large contact gap.
Capacitor bank	If switching transients set the design, the reduced stress allows tighter margins, higher steady-state utilisation, and lower unit-failure rates — better availability and lower recertification cost.
Power system	Better power quality: smaller switching overvoltages and travelling waves, less transient coupling into control and protection wiring, and a limited transient rise of the earth-mat potential — safer for personnel.

Controlled closing

For capacitor-bank energisation the aim is to close each pole when the voltage across that pole is close to zero, so the capacitor is not suddenly forced to a very different voltage. That voltage zero-crossing is the optimum making instant, where the source-to-capacitor voltage difference — and hence the inrush — is smallest. The controller monitors the source voltage and delays a random close command by an amount computed from the breaker’s mechanical closing time and its pre-arcing behaviour. As the contacts approach, the gap’s withstand falls along a slope called the rate of decrease of dielectric strength (RDDS) — how fast the gap loses its withstand as the contacts approach or separate — so a pre-arc strikes slightly before the contacts touch. Because real mechanisms scatter — the actual contact touch or separation time varies by a few milliseconds from the commanded time, enough to move the operation off the intended point on the wave — the controller aims for a point just after the zero-crossing to absorb both mechanical and dielectric scatter; a breaker for controlled capacitor switching should have a mechanical scatter below ±1 ms, preferably below ±0.5 ms.

Controlled opening

On opening, the controller monitors the current and targets a chosen arcing time so the contacts are as far apart as practical when the recovery voltage peaks — giving the gap the best chance to withstand it. The recovery voltage rises as a 1 − cosine to about 2 pu (the trapped charge holds while the source swings to the opposite polarity). The worst case is a controller failure that leaves a zero arcing time: the gap is then smallest exactly when the recovery voltage is highest, maximising the restrike probability. The bank meanwhile self-discharges slowly — only to about 30–40% of its initial voltage after 1–2 minutes — and is treated as fully discharged after five discharge time constants.

Cost

The premium for equipping a breaker with controlled switching falls sharply with rated voltage, because the controller cost is fixed while the breaker cost rises. It should be weighed against the reduced breaker stress, the reduced arrester duty, the lower capacitor-bank stress and the improved power-quality performance it buys.

Table 2 — Cost of a controlled-switching breaker relative to a standard one.

Rated Voltage	Cost Ratio
72.5–170 kV	about 150–160%
245 kV	about 120–130%
362 kV and above	about 105–110%
Where instrument transformers must be added	up to about 170–180%

The ratios are lower again for gas-insulated substations, where the switchgear itself costs more. CIGRE WG A3.35 has studied controlled closing and opening of both grounded and ungrounded banks in detail.

Section 3

Pre-insertion Resistors

A pre-insertion resistor is inserted briefly during closing: it limits the first charging current and damps the initial voltage transient, and after a short time the main contacts bypass it so it does not stay in the normal current path. In more detail, a pre-insertion resistor (PIR, or closing resistor) is built into the breaker, in parallel with the interrupter. On closing, an auxiliary switch closes first, so the bank is energised through the resistor at the reduced voltage left after the drop across it — limiting the starting inrush. After a set insertion time, the interrupter shorts out the resistor and the auxiliary switch. The electrical insertion time runs from the pre-arc in the auxiliary switch to the pre-arc in the interrupter and is typically 10 to 20 ms. The resistor is chosen to minimise both inrush and overvoltage; the optimum is roughly 20% of the bank’s impedance, and the effect can be sharpened further with controlled switching. The resistor is not chosen by ohmic value alone: its insertion time, energy absorption, thermal recovery and duty for repeated switching or reclosing must all be checked.

A pre-insertion resistor mainly helps during energisation; it does not normally mitigate opening or restrike behaviour unless the design adds a specific opening function. The resistor has no role on opening — the auxiliary switch, having only a closing capability, always opens before the interrupter. Reliability is high; only the number of successive closings is limited by the resistor’s thermal capacity. A PIR raises the breaker’s asset cost by up to about 30% but adds essentially no maintenance over its life. (A few designs use a pre-insertion reactor for the same effect.)

Section 4

Current-Limiting Reactors

A current-limiting reactor increases the inductance in the switching path, which reduces the peak inrush or outrush current and lowers the current’s rate of rise — though the added inductance can also shift the circuit’s natural frequency, raise the transient recovery voltage or create resonance concerns. A current-limiting reactor (CLI, also called a transient-limiting inductor, TLI) is a fixed series inductor permanently in each phase of the bank — unlike a PIR, which is only in circuit during closing. Its frequency-dependent impedance is large during the fast inrush and negligible at power frequency, so it caps the inrush, the outrush when a breaker closes into a nearby fault, and the discharge current of a restrike, while also cutting the \(di/dt\) and the transient earth-rise. The reactor’s position — system side or capacitor side of the breaker, or on the neutral — does not matter to its effect. Typical sizes are a few hundred microhenries for inrush limiting and 0.5 to 2.0 mH for outrush limiting.

A CLI turns the bank into a crude filter (see the APS note on switching of harmonic filter banks); the resulting tuning is high — above the 7th harmonic — so de-energising overvoltages are negligible. The main caveat is a transient recovery voltage risk: if a fault leaves the reactor as most of the fault impedance with little capacitance between breaker and reactor, the TRV imposed on the breaker can exceed its standard rating. The other costs are the reactor’s own footprint, foundation and insulators, rising with system voltage above about 245 kV. Because a current-limiting reactor — unlike a pre-insertion resistor — stays permanently in service, its losses, voltage drop, insulation duty and any resonance effects must all be acceptable during normal operation, not just at switching.

Section 5

Solid-State Switches

Solid-state switches operate very fast and can control current without mechanical contact scatter, but they carry continuous losses, cooling and leakage penalties, limited voltage ratings and high cost — so they are not usually the first choice for conventional HV capacitor-bank switching. They turn current on and off in semiconductors rather than with moving contacts, acting within microseconds and freeing the switching from mechanical limits — and they can control impedance, mimicking a resistor or reactor. But they carry two intrinsic penalties: a finite on-state resistance that dissipates far more than a metallic contact (needing active cooling), and a high-impedance “open” state that still passes a leakage current, so a solid-state switch is never a true disconnector.

They are also overload-limited: a thyristor can carry a surge for only one current half-cycle before it must be turned off at the next zero, and IGBTs/IGCTs handle even less (though they can interrupt a developing short circuit if the fault is detected fast enough). Voltage ratings are low — roughly 4 kV for IGBT/IGCT and up to about 12 kV for SCRs, with silicon-carbide devices now pushing higher — so power-system voltages need many series-connected modules, filling the volume of a dead-tank breaker or a shipping container. With current ratings from a few hundred amperes to about 4 kA and the cost of cooling and protection, a pure solid-state switch rarely pays: for the same outlay a FACTS device (a thyristor-switched capacitor or STATCOM) gives flexible, continuous control of reactive power, not just on/off switching. Put another way, where continuous dynamic reactive-power control is needed a FACTS is more appropriate than a solid-state switch applied only as a capacitor-switching mitigation device. This is why solid-state devices are common as converters and FACTS but very rare as plain capacitor switches.

Section 6

The Hybrid Diode Switch

A hybrid diode switch combines mechanical contacts with diodes so that the current flow is controlled during the critical making and breaking instants, the aim being to avoid pre-strikes and restrikes while keeping normal-service losses low. It combines the best of both worlds: it exploits the fact that a diode naturally changes state at a voltage zero-crossing, and pairs anti-parallel diode stacks with a fast mechanical switch: the mechanical contacts carry the rated current and provide the open-gap isolation, while the diodes conduct only during the make and break. Current is commutated from the contacts to the diodes on opening and back on closing, all within a fraction of a half-cycle, using a servomotor-driven common shaft with a built-in inter-phase offset and closed-loop (resolver) feedback.

The result is a smooth switching event free of pre-ignitions, re-ignitions and restrikes, which is kind to the capacitors: the bank energises with almost no inrush regardless of its charge, and the inrush never exceeds about twice the nominal capacitive current. Power loss is that of an ordinary mechanical switch (the contacts always operate at no load), so a large number of daily operations is feasible, and the diodes need only be rated for the brief make/break conduction, not the full impulse voltage. The design is practical up to about 38 kV; the closing/opening instant was held to a spread of just 220 µs. Its limitations: the motion must be precisely controlled — a control failure can damage the switch — and it cannot close into a full short circuit without much larger diodes. In short, the hybrid diode switch can sharply reduce inrush and restrike-related stress, but its use is bounded by voltage rating, control requirements, fault-making capability and the availability of suitable commercial equipment.

Section 7

Choosing Between the Options

The comparison below should be read as an application guide, not a ranking: the best choice depends on which stress dominates the project. The five families trade cost, complexity and what they actually limit. Controlled switching and the hybrid diode switch reduce the transient by controlling the switching instant or the current path; pre-insertion resistors and current-limiting reactors reduce it once it has started, by adding damping or impedance; and solid-state switching is reserved for where continuous control is also wanted.

Table 3 — The five mitigation families at a glance.

Device	What It Limits	In Circuit	Notes
Controlled (point-on-wave) switching	Inrush, overvoltage and restrike	Timing only	Needs single-pole operation and low scatter (<±1 ms); cost premium 105–160% by voltage.
Pre-insertion resistor	Inrush and overvoltage	During closing only	R ≈ 20% of bank impedance; ~10–20 ms; +30% cost; no opening role.
Current-limiting reactor	Inrush, outrush and restrike discharge	Permanent	Hundreds µH (inrush) to 0.5–2 mH (outrush); watch the raised TRV.
Solid-state switch	Inrush, overvoltage and restrike transients — fast and controllable, subject to device rating and control strategy	Permanent	Losses, cooling, leakage, low voltage rating; rarely economic versus a FACTS.
Hybrid diode switch	Inrush (to ≤2×), all strikes	Diodes only at make/break	Restrike-free, low loss; up to ~38 kV; needs precise control, no fault-making.

Commercially, pure solid-state capacitor switches remain rare. A few solid-state current or transient limiters do exist — for example a solid-state current limiter (series IGBTs/SCRs), and a solid-state transient limiter that suppresses inrush with a DC reactor and a varistor before a thyristor bypasses the reactor for load current. The hybrid diode switch described above is a different, part-mechanical approach (rated, for one device, at 17.5 kV / 630 A at 50 Hz). Which of these is practical depends on the voltage level, the losses, the cooling, the cost and the availability of suitable commercial equipment.

Section 8

Common Mistakes

Most errors here come from choosing a mitigation device before identifying the dominant stress, or from comparing the options on initial cost alone rather than on the transient each one actually limits and the penalty it brings.

Common mistakes to avoid

Specifying controlled switching on a gang-operated three-pole breaker, where the poles cannot be timed independently and the benefit is largely lost.
Ignoring the breaker’s mechanical scatter — controlled switching needs <±1 ms (ideally <±0.5 ms), and the target is set after the zero-crossing to absorb it.
Treating a pre-insertion resistor as an opening aid — it works on closing only, and successive closings are limited by its thermal capacity.
Fitting a current-limiting reactor without checking the raised TRV it can impose when a fault sits just beyond it.
Reaching for a solid-state switch as a simple on/off device — the losses, cooling, leakage and low voltage rating usually make a FACTS the better spend.
Expecting a hybrid diode switch to close into a fault, or running it with imprecise control — both risk damaging it.

Section 9

Key Points

Mitigation-device checklist

Capacitor switching is a frequent, heavy duty (often 1–2 operations/day, inrush to ~5 pu at kHz), so the breaker and bank benefit from deliberate transient mitigation.
Controlled switching is the mainstay, and the cleanest solution where the breaker offers single-pole operation, a reliable controller and low mechanical scatter (<±1 ms): close near a voltage zero (limiting inrush and overvoltage) and open at a chosen arcing time (lowering restrike probability).
Its cost premium falls with voltage — ~150–160% at 72.5–170 kV down to ~105–110% at 362 kV and above (more if instrument transformers must be added).
A pre-insertion resistor (R ≈ 20% of bank impedance, ~10–20 ms, +30% cost) is effective for closing transients only — its timing, energy and thermal recovery must suit the required switching sequence. A current-limiting reactor (hundreds µH to 2 mH) is permanent and also limits outrush and restrike discharge, but its TRV, resonance, voltage-drop and loss impacts must be checked.
Solid-state switches are fast and controllable but lossy, leaky and low-voltage; a FACTS is usually the better investment, so pure solid-state capacitor switches are rare.
The hybrid diode switch gives restrike-free, low-inrush (≤2×) switching up to ~38 kV with low loss and many operations — at the price of precise control and no fault-making capability.
Start by identifying the dominant stress — inrush, outrush, remote overvoltage, restrike risk, TRV, repeated-switching duty or power-quality impact — then match the mitigation to it: timing devices (controlled switching, hybrid diode) reduce the transient at source, passive limiters (PIR, reactor) cap its magnitude, and confirm the choice, especially any reactor TRV, with an EMT study.

Section 10

References and Further Reading

The standards and technical brochures behind this guide.

CIGRE, Shunt Capacitor Switching in Distribution and Transmission Systems, Technical Brochure 817. Paris, France: CIGRE, 2020.
IEC 62271-100:2021, High-Voltage Switchgear and Controlgear – Part 100: Alternating-Current Circuit-Breakers. Geneva, Switzerland: International Electrotechnical Commission, 2021.
IEEE Std C37.012-2022, IEEE Guide for the Application of Capacitive Current Switching for AC High-Voltage Circuit Breakers Above 1000 V. New York, NY, USA: IEEE, 2022.
J. C. Das, Transients in Electrical Systems: Analysis, Recognition, and Mitigation. New York, NY, USA: McGraw-Hill, 2010.

Part 8 Reading now

Devices for Limiting Capacitor-Switching Transients

Controlled switching, pre-insertion resistors, current-limiting reactors, solid-state and hybrid diode switches — how each limits the transient, and what it costs.

Series progress 8 of 8

All eight parts

Explore more technical guides

You have reached the end of the capacitive-current-switching series. Browse the full APS technical library for more power-system study guides.