Generator Breaker Protection Coordination: Damage Curves, Settings and Selective Tripping


A standby board is usually coordinated against the utility source and then spends its life running on a generator whose fault current is smaller, slower to establish and falling from the first cycle. The result is rarely a board that refuses to trip. It is a board that trips in the wrong place, or holds a fault long enough to damage the cable the breaker was meant to protect.

Two sources, two coordination problems

Every standby board has at least two fault levels. The utility connection is stiff: its fault level is set by the network and changes little with the load on site. The generator is not stiff. Its fault current comes from the alternator, is limited by the subtransient and transient reactance, and falls away within a few cycles as the field flux decays. A coordination study that uses only the utility case is therefore half a study.

The usual symptom is selective failure in one direction. A feeder breaker that clears a fault cleanly on mains may trip more slowly on the generator, because the current is lower and the element sits further down its curve. Or the generator breaker trips first, taking the whole board dark for a fault that a feeder should have isolated. Neither outcome is a breaker fault. Both are setting faults.

There is a second, less obvious problem. The generator contribution depends on excitation and on the type of excitation system. A set with a fast, high-ceiling exciter holds current longer than a set with a slower system, and the difference matters at the time delays used for short-time coordination. That information sits in the manufacturer’s decrement data, not on the nameplate.

Coordination therefore starts with data, not with settings. The list below is the minimum needed before a curve is drawn.

Information needed before a coordination study on a generator-backed board. Fault levels, decrement data and device characteristics are determined by the manufacturer’s data and the project specification.
Item Why it changes the answer Source
Single-line diagram and device ratings The study is only as good as the topology it is built on As-built drawings
Utility fault level at the point of connection Sets the maximum current the mains breaker must clear Utility data and project specification
Generator subtransient, transient and synchronous reactance Sets the current the set contributes and how fast it decays Generator datasheet
Generator decrement curve Shows the current available at the moment each element acts Generator datasheet
Breaker trip curves and setting ranges Determines what can actually be set on the device Breaker manufacturer data
Cable and busbar short-circuit withstand Forms the boundary the settings must stay below Project design data
CT ratios, classes and burdens Errors and saturation shift the effective setting Protection design
ATS transition mode and neutral arrangement Decides whether both sources can feed a fault at once Project specification

The damage curve is a boundary, not a target

A damage curve describes how long a component can carry a given current before it is damaged. Cable, busbar, transformer, generator and breaker each have one. The coordination plot places the protective device curve below the damage curve, so that the device opens before the protected item is degraded.

The common mistake is to set the device close to the damage curve to gain margin elsewhere. That margin is then consumed by the things that are not on the plot: current transformer error, the DC offset of the first cycles, ambient temperature in a room that reaches 40 °C in summer, and the tolerance band of the trip unit. The gap between the setting and the damage curve is determined by the manufacturer’s data and the project specification, not by a rule of thumb.

On a generator-backed board the generator belongs on the plot too. The stator winding, the damper winding and the exciter each have limits, and the breaker that protects the generator has to sit below them. The same breaker must also ride through the start of the largest motor, so the window between the start and the generator damage limit can be narrow. If the window closes, the answer is a different starting method, not a higher setting.

The cable damage curve is often the binding constraint on a long feeder. Its shape is steep at high current and flatter at low current, so a long-time element set for load can still sit above the cable continuous rating. Both ends of the curve have to be checked.

Curves on the coordination plot and what each one protects. The shape of each curve is taken from the component manufacturer’s data.
Curve Protects Read with
Generator decrement curve The alternator winding and its excitation limit The breaker that carries the generator current
Cable damage curve The insulation and conductor of each feeder The feeder breaker and the device above it
Breaker trip curve The device itself and the circuit it clears The trip unit data for the installed device
Transformer damage curve The winding, where a step-up or step-down is on the board The breaker on each side of the transformer
Relay characteristic The protected zone, including the earth path The CT ratio and the fault study
Busbar withstand rating The board and its joints The main breaker’s instantaneous element

Fault current from a generator is not a constant

A fault on a generator bus draws current that changes shape over time. In the first cycles the current is limited by the subtransient reactance and is at its highest. It then falls through the transient period as the flux in the damper winding decays, and settles at a steady value set by the synchronous reactance and the excitation ceiling. On some sets the steady value is a small fraction of the first-cycle value.

This matters because the time delay of the upstream device may be longer than the decay. A short-time element set to coordinate with a downstream breaker may be waiting while the current falls below its pickup, so it never operates, or operates late. The decrement curve is the only reliable way to know what current is available at the moment the element is asked to act.

The practical consequence is that instantaneous elements must be set against the minimum fault current at the end of the zone, not the maximum at the source. The minimum occurs at the remote end of the longest run, and on a generator source it is also the current after the decay. Both effects point the same way: the setting has to be low enough to see the smallest fault and high enough to ignore the largest start.

Voltage-dependent overcurrent is one way to keep an element sensitive during a fault while riding through a motor start, because the voltage collapse during the fault allows a lower effective pickup. Whether it is needed, and how it is set, follows from the manufacturer’s data and the project specification.

Indicative time bands in a generator fault. The actual decrement is taken from the generator manufacturer’s data, and the protection response is agreed in the project specification.
Time band What the current is doing Protection to consider
First cycles Highest current, limited by the subtransient reactance Instantaneous element and breaker instantaneous trip
Up to about one second Current falls as the transient flux decays Short-time delay element
After a few seconds Current settles at the value set by synchronous reactance and excitation Long-time element or voltage-dependent overcurrent
Throughout the fault Depends on the excitation ceiling and the exciter type Check the decrement curve at each element’s trip time

Setting the window: above the start, below the fault

The classic window has two edges. The upper edge is the largest current the board is expected to carry without a fault: the running load plus the starting current of the largest motor, with the starting contribution taken from the motor datasheet. The lower edge is the smallest fault current the device must clear, taken at the end of the protected zone and after the generator decay.

If the two edges cross, the board cannot be coordinated with the devices as selected. The options are to change the device, change the starting method, split the board, or accept a different protection philosophy with a documented rationale. Raising the setting to stop a trip during starting moves the upper edge up and makes the crossing worse, not better.

Time is the second dimension. The downstream device must clear before the upstream device reaches its trip time, with a margin that covers the breaker opening time, the relay timing tolerance and the CT error. The margin is set by the project specification; the curves provide the arithmetic.

Earth fault is usually the element left until last and then set by habit. Its pickup has to be above the standing leakage of the installation and below the minimum earth fault current at the end of the longest run. On a generator source the earth fault current can be limited deliberately, so the second number is a design output, not an assumption.

Setting window worksheet for each element. The values that fill the window come from the motor datasheet, the generator decrement data and the fault study; the margins are agreed in the project specification.
Element Must be above Must be below Decided by
Long-time pickup Generator rated current with the agreed margin, and the largest sustained load Cable ampacity and breaker rating Manufacturer data and project specification
Long-time delay The duration of the load peaks The cable damage curve at that current Project specification
Short-time pickup The largest motor starting current The minimum fault current at the end of the zone Motor datasheet and fault study
Short-time delay The downstream device clearing time plus the agreed margin The upstream delay, and the generator decrement at that time Coordination study
Instantaneous The asymmetric peak of the largest motor start The minimum fault current at the remote end Motor datasheet and fault study
Earth fault The standing earth leakage The minimum earth fault current at the end of the run Earth loop study and project specification

Transfer, ATS and the brief parallel window

The ATS adds a third case to the study. When the mains fails, the generator breaker closes only after the set is within its voltage and frequency window, and the load transfers in one step or in blocks. If the transition is closed, with both sources briefly paralleled, a fault during that window is fed from both directions at once and the fault level is the sum. The case is short, but it is the worst case for the downstream device.

Control power deserves its own line in the study. The protection relays and the breaker trip coils need a supply while the board is transferring. A UPS that covers the seconds between the mains interruption and the generator carrying load keeps that supply alive; without it, a fault during the gap may have no trip energy at all.

Neutral arrangement changes the earth fault path. A separately derived system with a switched neutral has a different earth fault current from a solidly connected one, and the settings have to match the arrangement that is actually built. It is worth confirming at commissioning rather than at the first fault.

The generator breaker also has to coordinate with the ATS mechanism. A transfer that closes the generator breaker onto a fault, or onto a load that is too large for the set, will trip the generator breaker and may leave the ATS in an intermediate state. The sequence should be tested with the largest load connected, not only with a light board.

Source conditions that change the coordination answer on a standby board. Fault levels are calculated for each case in the study, and the sequence is confirmed against the project specification.
Condition Fault level seen by the board What to check
Mains supply only Highest, set by the network Mains breaker and downstream devices coordinate at the utility level
Generator only, at rated output Lower, and decaying after the first cycles The downstream device still trips before the generator breaker
Closed transition, both sources closed The sum of both contributions The brief parallel window, and every device that sees the sum
Generator at 10% to 20% light load Lowest contribution, with the least stable voltage Whether the smallest fault is still detected
Cold set in a 40 °C room Current per the decrement data, with reduced thermal margin Breaker and cable derating for the installed ambient

Verification: injection, trip tests and the annual load test

A coordination study is a prediction. Commissioning is where the prediction is checked. Secondary injection confirms that the relay behaves as its settings say; primary injection or a trip test confirms the CT ratio, the wiring polarity and the trip circuit; a breaker timing test confirms the mechanism. None of these replaces the others.

The annual load test at 50% to 100% of rating is also a protection check, because it is the only time the board carries a current close to its rating with the generator as the source. The 10% to 20% light-load band is the opposite end: it is where voltage and frequency are least stable and where a marginal setting is most likely to show itself. Record the current in each phase, the voltages, the frequency and the breaker temperatures.

Settings drift for ordinary reasons: a trip unit is replaced, a CT is changed, a feeder is extended, a motor is added, a generator is swapped. Each of those is a reason to re-run the study, not only to adjust one number. The as-built settings sheet should carry a revision number and the date of the last study.

Ambient temperature is part of the check. A generator room that reaches 40 °C in summer reduces the thermal margin of both the breaker and the cable, and the manufacturer’s derating data applies to the device, not to the room. The setting that was correct at a cooler ambient may not be correct in the hot month.

Commissioning and annual verification of a generator-backed protection scheme. Intervals and acceptance margins are determined by the manufacturer’s data and the project specification.
Check Method Record
Settings as designed Compare as-found values with the study Settings sheet with date and revision
Relay characteristic Secondary injection at each point on the curve Injection test sheet
CT and trip circuit Primary injection or a trip test Ratio, polarity and trip record
Breaker opening time Timing test at each element Timing record for each device
ATS transfer and return Simulate mains loss with the largest load connected Sequence record with times
Annual load test 50% to 100% of rating, plus the 10% to 20% light-load band Load, duration, fuel and temperatures
Settings review Re-run after load growth, a motor addition or a generator change Revision note and new study
Can I reuse the utility coordination study for the generator case?

Only the topology and the downstream device data carry over. The fault level, the source impedance and the decrement all change, and the generator case usually sets the minimum fault current. Run both cases and coordinate each device against the worse of the two at every point on the plot.

Why does the generator breaker trip before the feeder breaker?

Two causes are common. The feeder short-time element may be waiting for a current that has already decayed below its pickup, or the generator breaker instantaneous element may be set below the fault current the feeder is expected to clear. Compare the decrement curve with the feeder clearing time before changing any setting.

Does a larger generator solve a coordination problem?

It raises the fault current, which can help a downstream device see the fault, but it also raises the starting current and the damage limit at the generator. It is a system change with several effects, not a setting change. The study has to be re-run with the new decrement data before the larger set is ordered.

Do I need a damage curve for the cable as well as the breaker?

Yes. The cable curve is often the binding constraint on a long feeder, and it has to be checked at both the high-current end and the continuous end. A long-time element set for the load can sit above the cable continuous rating even when the short-circuit check passes.

How do I coordinate with a soft starter or a drive?

The starter reduces the starting current, which widens the window, but it adds harmonics, a bypass contactor and its own protection logic. Coordinate the bypass and the drive input protection with the feeder breaker, and check the harmonic current against the generator derating discussed in the harmonic mitigation guidance.

How often should the protection settings be reviewed?

At commissioning, at the annual load test, and after any change to the load, the motors, the generator or the ATS. The review is a study, not a visual check. Keep the as-built settings sheet with the study revision so the next engineer can see what changed and why.

Check the window before the next fault

Send the single-line diagram, the utility fault level, the generator decrement data and the trip curves for the devices on the board. The review will set out the two fault levels, the window for each element, the transfer cases to test and the records to keep at handover.

Request a quote WhatsApp the sales desk

KVA Grid sales engineering desk. Reused site figures: the annual load test at 50 % to 100 % of rating and the 10 % to 20 % light-load band from the maintenance schedule; 500 kVA at 101/80/60 L/h, 1000 kVA about 200 L/h and about 4800 L over 24 hours, and 0.24 to 0.29 L/kWh from fuel consumption; 40 °C ambient from hot and humid climates; the PUE 1.4 base-load example and the UPS seconds between mains loss and set load from generator and UPS interaction; starting current and dip checks from motor starting; transfer sequencing from ATS transfer testing; parallel operation from paralleling and droop protection; load bank work from load bank testing; handover records from the commissioning checklist. Breaker trip curves, damage curves, generator decrement data, setting margins, CT performance and test intervals are determined by the manufacturer’s data and the project specification. Reviewed 9 November 2026.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top
WhatsApp