A genset short circuit level comes from the sub-transient reactance of the alternator, and it decays within a few cycles of the fault. Tesla Power Equipment sets out the calculation below, then shows how the figure changes at the far end of a feeder, so a switchgear quotation can be checked before the order.
- Sub-transient reactance first, because it sets the first-cycle current and no other number matters more.
- The decay curve second, since a device rated on the first cycle covers every later point on it.
- Cable impedance third, which pulls the figure down at the far end of a long run.
- Motor contribution last, kept beside the frames we build for export.
What does a genset short circuit level decide for a site?
It decides the rating of every device downstream of the set. A board rated below the fault level of the alternator is a board that fails on the first fault, and the failure is mechanical rather than electrical. The contacts weld, the arc escapes the chute, and the panel becomes the fault it was meant to clear.
A 625kVA frame at 400V carries 902A at full load, since 625,000 ÷ (1.732 × 400) = 902A. With 11% sub-transient reactance the same frame delivers 625 ÷ 0.11 = 5,682 times its rating in kVA, which at 400V is 8,203A. Every device on that board has to interrupt the figure, and the busbar has to hold it without flexing. The figure also sets the arc flash boundary, so a low reactance alternator pushes that boundary further back.
How is the first-cycle fault current calculated?
Divide the rating in kVA by the per-unit sub-transient reactance, then convert to current at the line voltage. A 625kVA frame with 0.11 per unit gives 5,682 units of fault duty on the same scale, and that current works out at 8,203A. That figure is the symmetrical root mean square current at the instant of the fault.
The peak asymmetrical current is higher, because the direct current offset adds to it. On a set with an X/R ratio near 15 the peak reaches about 2.5 times the symmetrical figure, so 20.5kA. A device is rated on the symmetrical value for breaking and on the peak for making, and the two appear as separate lines on the certificate. A buyer who reads only the kVA misses both figures, and both decide the switchgear cost.
Why does the current fall after the first cycle?
It falls because the rotor field cannot hold its flux against the armature reaction. The sub-transient reactance applies for the first 20ms to 50ms, the transient reactance for the next few hundred milliseconds, and the synchronous reactance after that. Each stage carries less current than the one before it, and each stage has its own data sheet symbol.
On a 625kVA frame the first cycle reads 8.2kA, the value at 100ms reads about 4.9kA, and the value at 500ms reads about 2.9kA. Those three follow from the machine constants rather than from a rule of thumb, and the alternator test sheet states them as Xd, Xd prime and Xd double prime. The three reactance values belong on the alternator test sheet, and a quotation that omits them cannot be checked against a second one.
| Point in time | Reactance in use | Current at 625kVA | What it sizes |
|---|---|---|---|
| First cycle, 20ms | Sub-transient | 8.2kA | Device breaking capacity and busbar bracing |
| 100ms | Transient | 4.9kA | Short-time delay and cable withstand |
| 500ms | Transient, falling | 2.9kA | The point where protection should have cleared |
| Steady state | Synchronous | 1.6kA | Sustained current into a stuck contactor |
How does 100ms change the device choice?
It changes it where the device has a short-time delay, which most main breakers do. A device set to hold 6kA for 100ms sees 4.9kA at that instant and holds, which is what the delay is for. A device set to hold 9kA for the same 100ms sees the same 4.9kA and holds too, so the extra setting buys selectivity rather than protection.
A genset short circuit level in a plant room with two sets needs the sum rather than either machine. The 100ms figure also sizes the cable. A 240mm² copper conductor withstands about 8kA for one second, so the same conductor holds 4.9kA for about 2.6s, since the withstand falls with the square root of time. Most genset manufacturers leave that check to the installer, and Chinese genset suppliers usually quote a device rating without the decay curve behind it. The check takes five minutes and prevents a welded contactor.
What does cable impedance do to the figure?
The cable impedance pulls the current down, and on a long feeder it pulls the figure down far enough to change the device. A 240mm² cable has about 0.075Ω per km of resistance and 0.08Ω per km of reactance, so a 50m run adds roughly 0.0055Ω to the loop. At 400V that impedance limits the current well below the terminal value.
The limit swaps hands as the run grows. At the terminals the alternator reactance decides the current. At about 150m of 240mm² the cable impedance and the alternator impedance are equal, and beyond 250m the cable decides and the value falls toward 4kA. A device chosen on the terminal figure is then overrated for the board it protects. That swap is why a device chosen for the terminals can be wrong at the board.
- At the terminals, the alternator reactance decides the current.
- At 150m of 240mm², the two impedances are equal and the current sits near 5.5kA.
- Beyond 250m, the cable decides, and the value falls toward 4kA.
How much do motors add to the fault current?
Motors add current for the first few cycles, because a spinning motor behaves as a generator until its field collapses. A motor adds roughly four to six times its rated current at the instant of the fault, and the contribution decays within 100ms to 200ms. An 80kW motor at 0.8 power factor draws 144A at 400V, so it adds about 700A for that short window, and six of them add 4.2kA.
The addition matters at the terminals and not at the far end, because cable impedance damps the motor contribution first. A panel with six motors of that size adds 4.8kA to the first cycle, which lifts 8.2kA to 13kA. A device rated at 10kA then fails the check, and the answer is a higher rating rather than a smaller set. The contribution is a first-cycle problem, and the device rating is where it shows.
Where does the calculation go wrong on an export order?
It goes wrong at the reactance figure. A buyer takes the value from a similar set in an older project, and the alternator on the new order has a different winding. The gap between 11% and 14% moves the fault level from 8.2kA to 6.4kA, which is enough to change the device rating and the busbar bracing.
The second error is a fault level quoted at the alternator terminals and applied to a board 80m away. The board sees less current, so a device chosen on the terminal figure is overrated, and the overrating is harmless until the site adds a second set and the figure doubles. Our export desk asks for the single line diagram before it quotes a device. Both errors are cheap to avoid and expensive to find after a panel has failed.
What changes when two sets run in parallel?
The fault current adds, though not in a straight line. Two identical 625kVA sets in parallel deliver about twice the current of one into a bolted fault at the common busbar, so 16.4kA rather than 8.2kA. The busbar, the devices and the cable all move up a bracket, and the bracket is not optional.
The addition is close to arithmetic because the two sources share the fault through their own reactances in parallel. A 625kVA set paired with a 1000kVA set gives a sum weighted by rating, so the smaller machine contributes about 38% of the total. Yangzhou Tesla Power calculates that sum for every parallel order and prints it on the drawing. The drawing states the figure at the common busbar rather than at either alternator, because that is where the devices sit.
What should the offer state about fault level?
Four figures. The sub-transient reactance as a percentage, the first-cycle symmetrical current at the terminals, the peak asymmetrical current, and the current at 100ms. A quotation that states only the kVA leaves the buyer to assume a reactance, and the assumption is usually optimistic by two or three points.
The ISO 8528-5 transient limits set the voltage dip against which the fault current is measured, and the genset CE marking requirements put the device rating inside the technical file. Ask for the four figures before comparing two quotations, because the cheaper set often carries the higher reactance. The four figures are not a formality. A buyer who compares two sets on rating alone compares two machines and not two protection schemes, and the difference appears at the first fault rather than at the invoice. Ask for the reactance in writing, and treat a supplier who declines as a supplier with a reason.
Which loads make the fault level irrelevant?
A set that feeds only resistive load through a fused isolator rarely sees a bolted fault at the board, so the fault level matters less there than the cable rating. A set that feeds a drive panel does see one, because a rectifier fails short and the fault sits at the terminals with nothing between it and the device. The reference condition is defined in the ISO 8528-1 rating definitions, which names the tolerance and the test point.
A genset short circuit level for a drive panel starts at the terminals, a genset short circuit level for standby duty starts at the far end of the feeder, and a lighting board starts at the cable. The distinction decides how much arithmetic the order needs. A lighting and heating panel takes the terminal figure and a 10kA device. A panel with drives and a 250m feeder takes the far-end figure, the motor contribution and the decay curve, and the answer changes twice before it settles. A drive panel usually carries its own breaker, so the fault level question moves to the drive rather than to the set, and the drive maker answers it. The same frame appears in the 120kW set on the WP6 platform, where the plate states the same basis.
What does the site keep for the next fault?
Three documents. The alternator test sheet with the reactance figures, the calculation sheet with the four currents, and the device label with its rating. A site that holds all three answers a question about a welded contactor in an hour rather than in a week of email and photographs. The same frame appears in the 1000kW container build for standby duty, where the plate states the same basis.
Tesla Power sends that pack with every set and keeps a copy in the order file. The same figures appear on the panel drawing, so a switchgear contractor prices against the numbers the alternator was ordered against. Send the single line diagram to our export desk and the reply comes back with the four currents filled in. The three documents are small enough to travel in a folder. A site that cannot find them rebuilds the calculation from the nameplate, and the nameplate carries the rating without the reactance that made the rating useful.
It is proved by a primary injection test at the board, or by a calculated value accepted against the alternator test sheet. An injection set applies a known current through the device and confirms that it trips at its setting, which proves the device rather than the source. The source figure comes from the machine constants. The same frame appears in the 400kW frame on the Weichai WP13 platform, where the plate states the same basis. Tesla Power reads that figure on the site sheet before the plate is printed.
A site without injection equipment uses the controller log. Most controllers record the current at the moment of a trip to within 5%, and that reading is a real fault rather than a calculation. We ask for that log whenever a site reports a device failure, because it settles in one line what a page of arithmetic debates. The log is worth more than the arithmetic once a fault has happened, because the fault measured the set rather than the sheet. The same frame appears in the 500kVA silent set with the ATS option, where the plate states the same basis.
Frequently asked questions about genset short circuit level
The value on the alternator test sheet for that order, which usually sits between 10% and 14% for a four pole machine. A value taken from a similar set in an older project is a guess, and a guess on this line moves the device rating by a whole bracket. Tesla Power Equipment keeps the correction on the offer, so two quotations can be compared.
The per-unit reactance is close, and the current differs because voltage and rating change. A 625kVA frame at 400V and at 480V gives 8.2kA and 6.8kA respectively, so an export order to a 480V market needs its own calculation rather than a copy.
Until a device clears it, which should be under 0.4s on a 400V installation. If nothing clears it, the current falls to about 20% of the first-cycle value within two seconds and the alternator overheats rather than the cable. Tesla Power reads that figure on the site sheet before the plate is printed.
Usually, because the rating rises faster than the reactance. A 1000kVA frame at 12% gives 12kA where a 625kVA frame at 11% gives 8.2kA. A large frame with a deliberately high reactance can give less than a small one, which is why the test sheet matters more than the rating.
It can, and the margin is normal practice. A 10kA device on an 8.2kA fault holds a 22% margin, which covers a later second set or a motor contribution the original calculation missed. Tesla Power reads that figure on the site sheet before the plate is printed.
It sizes the busbar bracing and the making capacity of the device. A contactor that closes onto a fault sees the peak rather than the symmetrical value, and a busbar that flexes under the peak throws a cable into the next panel.
It falls slightly, because winding resistance rises with temperature while the reactance stays fixed. The change is small enough to ignore on a standby set and worth a recheck on a prime set that has run 30,000 hours.
The alternator test certificate with the reactance figures, signed at the works. A site that adds a second set later uses that certificate rather than the nameplate, since the plate carries the rating and not the reactance. Tesla Power reads that figure on the site sheet before the plate is printed.