Genset breaker sizing sets three things at once: the frame, the trip setting and the cable that carries the load. Tesla Power Equipment works the three together below, so a quotation states a device that protects the conductor rather than one that only looks correct on a schedule.
- Full load current first, because every later figure is a multiple of it and a guess here carries through the whole order.
- Frame against setting, since a 1000A frame with an 800A trip behaves differently from an 800A frame at its limit.
- Cable rating after derating, which decides whether the breaker protects the conductor or merely follows it.
- The fault level at the terminals, kept beside the frames we build for export.
What does genset breaker sizing decide for a site?
It decides whether a fault clears at the set or at the board, and it decides which of the two devices moves first. A breaker chosen on full load current alone trips on a motor start and holds through a real fault. A breaker chosen on fault current alone sits above the working load, so an overload heats the cable without moving the trip.
A single ratio ties the two together. Set the trip at 1.1 times full load and check that the cable still clears at that current after derating. Where the two disagree, the cable changes and not the trip, because the device exists to protect the conductor and only secondarily to protect the alternator. Tesla Power applies that order on every offer.
How is full load current calculated for a three phase set?
Divide the rating in volt amperes by root three times the line voltage. A 500kVA frame at 400V carries 722A, since 500,000 ÷ (1.732 × 400) = 722A. A 250kVA set at the same voltage carries 361A, and a 100kVA set carries 144A. At 480V the 500kVA frame carries 601A instead, which is why an export order names the voltage before the rating rather than after it.
Power factor does not change that current, because the alternator is rated in kVA. It changes the kW the engine has to deliver, which is a separate question about the prime mover. A 500kVA frame at 0.8 power factor carries 400kW, and the same frame at 0.9 carries 450kW, which the engine may not sustain for a week in a hot yard. A quotation that lists kVA alone hides that difference.
Which frame size fits a given alternator?
The frame is the current the device carries continuously, and the trip is the current at which it opens. A 500kVA set at 722A fits an 800A frame with the trip at 0.9, which gives 720A. The same set on a 1000A frame with the trip at 0.72 gives the same 720A and costs more for no gain in protection.
Most genset manufacturers fit the smallest frame that covers the current, because the device is a cost line rather than a design feature. Most genset factories in China publish that frame without the derating basis. That choice holds on a single set in a cool room. It becomes tight where the set carries 722A for eight hours and the panel sits at 45 ℃ inside a canopy, since the frame derates with panel temperature.
| Rating at 400V | Full load current | Frame and setting | Fault level at 12% Xd |
|---|---|---|---|
| 100kVA | 144A | 160A frame, trip at 0.9 | 1.2kA |
| 250kVA | 361A | 400A frame, trip at 0.9 | 3.0kA |
| 500kVA | 722A | 800A frame, trip at 0.9 | 6.0kA |
| 1000kVA | 1,443A | 1600A frame, trip at 0.9 | 12.0kA |
How does the trip setting relate to the frame?
The setting is a dial on the trip unit and it moves inside a range the frame allows. An 800A frame usually accepts 0.4 to 1.0 of its rating, so 320A to 800A. A site that adds load later can turn the dial up without changing the device, which is why the frame is chosen for the future and the setting for today.
The long time delay decides how long the device holds at 1.2 times the setting, and that delay is the figure most often left at its factory value. A setting of 720A holds 864A for anywhere from 2s to 30s depending on the curve. Tesla Power sets the delay at 8s on a standby build, because a motor start of 3s should not trip a board that carries the plant.
What breaking capacity does a set need?
The device has to interrupt the fault current the alternator delivers, and that figure comes from the sub-transient reactance. A 500kVA frame with 12% reactance delivers 41.7 times its rating into a bolted fault, which at 400V is 6,014A. The device therefore needs a rating above 6kA, and 10kA is the usual choice for margin.
The first cycle is the worst point. By 100ms the current sits near 60% of the first cycle value, and by 500ms it is closer to 35%. A device rated on the first cycle figure covers every later point on the curve, so the first cycle is the number to quote. A device chosen on the 500ms figure lets the first cycle through the contacts.
- 6kA on a 500kVA frame at 12% reactance, taken at the terminals.
- 3kA on the same frame with the fault at the far end of a 50m feeder, where the cable impedance adds to the loop.
- 12kA on a 1000kVA frame at the same reactance, which moves the device into another cost bracket.
A genset breaker sizing in a plant room follows the panel temperature rather than the yard average, and that single line moves a 250kVA order from a 3kA device to a 6kA one.
How does the cable change the breaker choice?
The breaker protects the cable, so the derated cable rating has to sit above the setting. A 240mm² copper four core cable carries about 480A in free air and about 380A in a duct. Two runs in parallel carry 960A and 760A respectively, and the 720A setting fits the duct figure with nothing left for a hot room. A third run buys margin, and it costs less than a new device.
Derating follows the installation rather than the catalogue. A cable at 45 ℃ ambient loses about 10% of its figure, a bank of six touching cables loses another 20%, and a length inside thermal insulation loses more again. The site planning notes set out the factors, and the schedule records the derated figure beside the setting so the two are read as one line. Where the derated figure falls below the setting, the cable changes.
Where does coordination between two devices go wrong?
Coordination goes wrong when the upstream device is slower than the downstream one at every current, and it goes wrong without a symptom. Two devices of the same make and curve coordinate by ratio, so a 400A downstream under an 800A upstream works when the settings follow. Two makes need the curves overlaid rather than compared by rating.
A generator set adds a case a mains board does not have. The fault current is limited by the alternator, so the current at the far end of a long feeder can fall short of the instantaneous band of either device. Then neither trips on the magnetic element, and the thermal element takes seconds over a fault that should have cleared in 0.1s. Yangzhou Tesla Power plots both curves before dispatch.
How is the choice proved before shipment?
Prove it with a primary injection test on the device and a secondary injection test on the trip unit. The first proves the current path and the mechanism, and the second proves the curve. A set that passes both leaves the works with a sheet naming the setting, the curve and the current applied at each stage. A buyer can read that sheet without an interpreter, because every line names a figure and a device.
We run both tests on every order above 250kVA, because the cost of a bench test is a fraction of a site visit after a nuisance trip. The sheet carries the serial number of the device, so a replacement in five years can be matched to the original curve rather than guessed at from a photograph of the label. The file also records the ambient at the test, since a curve proved at 25 ℃ is not the same claim at 45 ℃.
What should the order sheet state about the breaker?
Four lines. The frame rating, the trip setting with its curve reference, the breaking capacity in kA, and the cable size with its derating basis. A quotation that states only the frame leaves the other three to the installer, and the installer usually works from the cable on the shelf rather than the one the set needs.
The ISO 8528-1 rating classes fix the duty the current is calculated against, and the genset CE marking requirements place the device inside the technical file rather than on the plate. Buyers who ask for genset breaker sizing for a cable schedule get those four lines on the offer, and the fourth line is the one that changes the price.
Which mistakes come back from site?
Three mistakes come back often. The first is a frame chosen at the alternator rating with no margin, which trips on the first hot afternoon. The second is a setting raised at site to stop a nuisance trip, which then leaves the cable unprotected. The third is a device replaced with a higher rating after a failure, which moves the fault duty onto the board upstream and turns one failure into two.
Each has the same cure, which is a written setting and a curve reference that travel with the machine. A crew that knows the setting is 0.9 of 800A calls the factory before turning the dial. A crew that knows only that it tripped turns the dial and moves on, and the next fault finds a cable with no protection at all. The label costs nothing and prevents all three. The reference condition is defined in the NFPA 110 emergency power rules, which names the tolerance and the test point.
What makes the sizing auditable five years later?
A calculation sheet with four inputs: the alternator rating, the voltage, the cable schedule and the prospective fault current. Every later figure follows from those four, so a reader can rebuild the answer without the person who wrote it. That matters when a site adds load and nobody remembers why the frame was chosen, and it matters more when the person who chose it has left. The same frame appears in the 1000kW container build for standby duty, where the plate states the same basis.
Tesla Power keeps that sheet in the order file and sends a copy with the set. The same figure appears on a label inside the panel door, which is the version a technician reads at two in the morning. Conformity assessment for the assembly then checks that the label matches the file rather than the sales order, and the check takes a minute. A file with the four inputs answers a question about a trip in one call rather than in a week of email. The same frame appears in the 400kW frame on the Weichai WP13 platform, where the plate states the same basis.
Three records settle a switchgear dispute. The calculation sheet, the injection test report and the panel label. Where the three agree, a dispute about a trip becomes a question about the load rather than about the device. Where they disagree, the label is wrong and the file is right, and the correction is a label rather than a new breaker. The same frame appears in the 500kVA silent set with the ATS option, where the plate states the same basis.
Send the alternator rating, the voltage, the cable schedule and the prospective fault current to our export desk. The reply comes back as a device selection with the setting and the curve, and that reply is the document a switchgear contractor prices against rather than interprets.
Frequently asked questions about genset breaker sizing
At 100% for a set whose current is calculated, and at 125% for a continuous load that runs unattended. The 125% rule comes from continuous load practice rather than from alternator behaviour, and it costs a frame size on most standby builds.
Only partly. The alternator has its own thermal limit and its own short circuit withstand, and a breaker set for the cable sits above both on a long feeder. Protection against alternator overload comes from the controller, which reads current and winding temperature together. Tesla Power reads that figure on the site sheet before the plate is printed.
3kA at the terminals with 12% reactance, so a 6kA device covers it with margin. A device rated at 3kA sits exactly at the calculated fault and leaves nothing for a fault close to the terminal box.
It needs one where the controller has to open the device on a fault the breaker cannot see, such as low oil pressure or overfrequency. A shunt trip also lets an emergency stop drop the load without waiting for the thermal element.
Because the long time delay counts heat rather than current, and a device in a warm panel reaches its threshold earlier than the curve states at 40 ℃. The curve carries a temperature reference, and a panel at 55 ℃ moves the effective setting down by around 8%. Tesla Power reads that figure on the site sheet before the plate is printed.
On a label inside the panel door, with the frame, the setting, the curve and the date. A setting that exists only in a commissioning file is a setting nobody reads, and the next crew turns the dial to clear a trip.
Not for current, since 722A fits an 800A frame at 0.9. It needs the larger frame where the fault level exceeds the 800A device rating, or where the site plans to add more than 10% and wants the dial to move without a new device.
The current is the same, but the hours are not. A prime set holds 722A for most of the day, so the frame runs near its thermal limit and the panel temperature matters more. Tesla Power adds a size on prime duty for that reason.