The letter arrives from the utility, and it is two paragraphs long. The available fault current at the 13.8 kV service has been revised upward because a neighbouring substation was rebuilt and a transmission loop was closed. The metal-clad lineup in the plant’s main electrical room is rated 31.5 kA; the new number is 38 kA. Nothing in the plant changed, but as of that letter the switchgear is under-rated for the system it sits on. The reflexive answer is to price a replacement lineup. The less obvious answer is that fault duty can often be brought back under the existing rating by adding impedance to the system — and a current-limiting reactor is the device that does it.
What Is a Current-Limiting Reactor?
A current-limiting reactor is a series-connected inductance with no other job: an ohmic reactance at 60 Hz, a continuous current rating, short-time withstands, and an insulation level. Unlike the current-limiting fuse that shares part of its name — which interrupts the circuit to limit let-through — a reactor interrupts nothing and limits current continuously.
That is the whole trade-off in one sentence. The reactor’s benefit appears during a fault that might happen once in several years; its cost — voltage drop, losses, reactive consumption, floor area — is paid continuously for the 25 to 30 years the installation is in service.
Why Fault Current Grows Over Time
Available fault current is a property of the system the installation is connected to, not of the installation itself — and that system gets stronger over time. A rebuilt substation, reconductoring, a closed loop, or upstream generation all reduce source impedance, and the consequence lands on whoever owns the switchgear. None of it announces itself: the issue surfaces when the utility revises its fault data, a study is updated, or the installation is reviewed against the local code’s interrupting-rating requirements.
The same thing happens inside the fence. On an effectively infinite source, replacing a 1,500 kVA, 600 V transformer with a 2,500 kVA unit at the same 5.75% impedance raises calculated secondary fault current from approximately 25 kA to 42 kA. A real study also includes utility, conductor, motor and generator impedances.
The Switchgear Replacement Problem
Interrupting rating is not field-adjustable, so the conventional path is replacement, and replacing a medium-voltage lineup means cable re-termination, protection re-engineering, a plant outage of days, and lead times measured in quarters. A reactor inverts the problem: a new device in a new footprint, the existing gear untouched, a short tie-in outage, and capital at a fraction of a lineup.
The honest assessment: a reactor corrects one specific deficiency, and it is the right answer only when that deficiency is the only one. If the gear is obsolete, parts are unavailable, or the binding constraint is the bus’s short-time withstand rather than the breakers’ interrupting rating, a reactor buys time without solving the problem. And if a significant share of the fault contribution comes from motors or generation already on the protected bus, a reactor upstream of that bus does not reduce their contribution at all.
How a Series Reactor Limits Fault Duty
Fault current at a bus is source voltage divided by the total impedance between source and fault. A reactor adds to the denominator — the same mechanism that makes a transformer’s percent impedance the dominant limit on secondary fault current, except purchased as a separate device. The arithmetic is direct: at 13.8 kV with 40 kA available, source impedance is 13,800 ÷ (√3 × 40,000), about 0.199 Ω. Bringing the bus below a 31.5 kA rating requires approximately 0.253 Ω. Because a medium-voltage source is strongly reactive (X/R commonly 10 to 30), the required 0.054 Ω adds very nearly arithmetically, and a 0.055 Ω reactor gets you there.
Two consequences catch people. First, diminishing returns: halving fault current requires doubling total impedance, so a 20% reduction is cheap and a 50% reduction is a different project. Second, reducing symmetrical RMS current is not the same as reducing the asymmetrical peak: an air-core coil’s own X/R is very high, so inserting it raises system X/R, lengthens the DC time constant and increases asymmetry — sizing from symmetrical current alone may leave the breaker’s close-and-latch capability exceeded. Final sizing uses the system’s complex impedance with all sources, and the asymmetrical-duty check per IEEE C37.010 and the equipment ratings.
Air-Core vs. Iron-Core Reactor Design
An air-core reactor has no magnetic circuit to saturate, so its reactance is the same at 1,200 A continuous as at 31,500 A of fault current — linearity at 25 times rated current is the requirement. The cost is that the flux has nowhere to go but the surrounding air: air-core coils carry a magnetic clearance envelope, and closed loops of reinforcing steel, structural members, conduit and ferrous hardware inside it all see induced current. The field consequence is a warm wall or a hot anchor bolt nobody can explain — and in a retrofit switch room, the envelope usually kills the layout before price does.
Gapped iron-core reactors contain the flux, making them smaller and easier to place indoors — though gap forces make them audibly louder — and they are the practical choice for most low-voltage work. The trade-off: the gapped core must remain linear to the full asymmetrical peak, because a core that saturates under fault conditions loses its reactance at precisely the moment it was bought to provide it.
Sizing a Current-Limiting Reactor
Size backward from the rating you are protecting, not forward from a target percentage. The reactor exists to bring four numbers within limits — breaker symmetrical interrupting capability, close-and-latch or momentary capability, bus and cable short-time withstand, and the site’s arc flash target — and the lowest governs. Placement follows from what needs protecting: an incomer reactor limits only what passes through that incomer, not motor or generator contributions downstream of it, while a bus-tie reactor limits each section’s contribution to a fault on the other.
Three inputs define the specification: available fault current with its X/R ratio; the target reduction converted to ohms, with a minimum tolerance; and system voltage, BIL and continuous duty, with the coil specified and tested per IEEE C57.16-2025.
Calculate the steady-state penalties before the reactor is specified. That 0.055 Ω coil carrying 1,200 A drops 66 V per phase in quadrature — roughly 0.44% of nominal at 0.85 power factor — and consumes about 238 kvar continuously, plus real losses that belong in the room’s heat-load calculation. Lower fault current also moves every downstream device to a slower point on its time-current curve, so the fuse selection that coordinated at the old level may clear more slowly at the new one. Arc flash incident energy is a product of current and time, so a bus can end up with higher incident energy after the reactor than before it. Coordination and arc flash studies are re-run as part of the reactor project, not afterward.
Common Applications
Utility Substations
Substations have the longest track record, usually with bus-tie reactors: both bus sections operate closed for reliability without each source adding its full contribution to a fault on the other. Feeder reactors are more of a compromise — the same reactance that limits fault current deepens the voltage dip during downstream faults and motor starts, making it a power quality question as much as a protection one. During a 31.5 kA fault, a 0.055 Ω coil holds roughly 1.7 kV per phase: the mechanism that supports the healthy bus is the same one that pulls down the faulted feeder.
Industrial Plants Adding Generation
Adding a synchronous machine is the most common self-inflicted fault duty problem: with subtransient reactance typically 12 to 20%, a generator contributes several kA, additive when paralleled with the utility, and a series reactor at the machine is often cheaper than re-rating the distribution system. Inverter-based resources are the opposite case — battery and solar inverters are current-limited by their controls to roughly 1.1 to 1.5 times rated current, so a study that models them like synchronous machines produces an expensive reactor for a problem that does not exist. Check how the source was modelled before accepting the recommendation.
Data Centres
Larger service transformers, paralleled sources in N+1 and 2N topologies, closed ties and rising rack density all push fault current upward on switchgear sized for the original load. Canadian facilities typically distribute at 600 V, where the main switchboard can already sit near the 65 kA and 100 kA plateaus low-voltage gear is built to. Reactors are less common at low voltage — a higher-impedance replacement transformer or current-limiting fuses is frequently the cheaper fix — but where the switchboard is sound and only the incoming duty has moved, a series reactor at the incomer is far less invasive than replacing a live data centre lineup.
Installation and Space Considerations
Take the magnetic clearance envelope from the manufacturer’s outline drawing and treat it as a hard constraint at layout. For air-core units it extends in all three axes, dictating non-magnetic hardware, reinforcing steel in open rather than closed loops, and distances to structural steel and enclosure walls — discovering it after the pad is poured is a costly sequence to reverse. Stacked single-phase coils save floor area but couple magnetically, so per-phase reactance differs unless the design compensates; horizontal arrangements are cleaner but need the space. Fault forces scale with the square of peak current, so the mounting steel and anchorage are part of the fault rating. Indoors, add the heat to the cooling load; outdoors, suit the insulators to pollution and icing and confirm clearances with the authority having jurisdiction.
Rex Power Magnetics Perspective
Rex Power Magnetics builds dry-type air-core reactors, so the specified reactance holds from continuous current through full asymmetrical fault current. Units are supplied with the ohmic value and tolerance, continuous rating, short-time withstands, loss data and the magnetic clearance envelope. Where fault current, X/R or the governing rating is uncertain, Rex’s engineering group supports specification review against the short-circuit and coordination study before the order is placed — including the asymmetrical and arc flash checks that reactor projects most often miss.
Conclusion
A current-limiting reactor solves an intermittent problem with a permanent device: the fault it was bought for may never occur, while the voltage drop, reactive consumption, losses, floor area and slower downstream protection are present every hour of a 25 to 30 year service life. The exchange is worth making when the switchgear remains serviceable, the required reduction is achievable without unacceptable penalties, and most fault contribution originates upstream of the reactor. As the required reduction grows, compare the penalties directly against replacement — a reactor corrects a rating problem; it does not renew a lineup.