Generator Current Unbalance: A Sequence Issue, Not AVR

Brian Holt9 min read
Other ManufacturerTroubleshootingWiring & Electrical
Licensed PE Working through this on a live machine? A Maine-licensed engineer can take it from here — included with IMD hardware, by the hour for everything else. Book an engineer

The engineer sees a negative-sequence trip while a 2 MW, 480 V diesel generator is paralleled with the utility. Relay records show approximately 440 A, 450 A, and 505 A even though bus voltage stays within a volt or two and phase separation remains 120° ±0.5°. Treat the current split as a sequence-network or measurement problem; a conventional common-field AVR cannot selectively force one stator phase.

Stop blaming the AVR and repeated resets

Do not start by changing power-factor, VAr, or voltage-control settings. Mechanical input establishes real-power transfer and excitation establishes the machine's aggregate reactive-power operating point. Both act on the three-phase machine, not on one phase independently.

A conventional AVR regulates the common rotor field. It cannot produce sustained phase-selective field forcing as the rotor turns. An excitation problem can move all three currents by changing internal generated voltage, but it does not normally explain one phase at 505 A while the others sit near 440 A and 450 A. Resetting the negative-sequence trip without finding the source exposes the rotor and stator to additional sequence-current heating.

Balanced terminal voltage also does not prove balanced current. While paralleled to a stiff utility source, generator current depends on the small voltage difference across the impedance between the generator's internal voltage and the grid. A small negative-sequence voltage or a phase-dependent impedance can therefore produce a much larger percentage current unbalance while the bus voltages still look nearly equal.

Observed symptom Likely branch Next reading or test
Relay and separate metering show the same phase-current pattern Real current unbalance or a shared CT problem Identify each instrument's CT source and compare with an independent primary-current measurement
Only one instrument shows the unbalance Relay input, wiring, scaling, or acquisition problem Secondary injection and channel-to-channel comparison
Current unbalance appears only while connected to the grid Grid negative-sequence voltage, transformer effect, or unequal interconnection impedance Record voltage and current phasors at both sides of each intervening element
Open-circuit generator voltage is unbalanced Generator winding, connection, sensing, or internal magnetic asymmetry Compare phase-to-phase and phase-to-neutral voltage with the machine isolated
Open-circuit voltage is balanced but loaded current is not CT circuit, conductor path, joint, transformer, or system interaction Inspect and electrically test each phase path

Check 1: Prove what actually tripped

Read the event report before disturbing settings. Confirm that the operated element was negative-sequence current and capture the pre-trip current and voltage phasors, not just RMS magnitudes. The installation reported a negative-sequence operation, and separate metering displayed the same current imbalance. Two sets of analog-to-digital converters reduce the chance of a single failed converter, but they do not eliminate a shared CT, shared terminal block, polarity error, or common wiring fault.

The arithmetic average of 440 A, 450 A, and 505 A is 465 A. The highest phase is 40 A, or about 8.6%, above that average; the total high-to-low spread is 65 A, or about 14.0% of the average. Those figures describe magnitude spread only. They are not the negative-sequence percentage.

Calculate negative-sequence current from synchronized phase-current phasors using the symmetrical-component relationship I2 = (Ia + a²Ib + aIc) / 3, where a represents a 120° rotation. Current magnitudes alone cannot supply I2 because their angles are missing. Compare the relay's calculated I2 with an independent instrument at the same instant and verify the protection pickup and delay against the approved protection study and generator capability data before changing either setting.

Check 2: Measure voltage at the current-transformer location

Use synchronized phase-to-neutral and phase-to-phase voltage phasors at the point associated with the relay CTs. A statement that the bus is “rock steady” describes time variation, not sequence balance. Record the negative-sequence voltage as well as the three magnitudes and angular separations.

The recorded voltage in this case deviated by no more than a volt or two, with angular separation within 0.5° of 120°. That is much better than a prior operating case in which phase voltages differed by more than 10 V on a 480 V base and phase separation ranged from below 116° to above 121°. It still does not close the grid-interaction branch: a stiff grid can hold generator terminal voltage close to its own voltage while circulating appreciable negative-sequence current through the generator's relatively low sequence impedance.

If negative-sequence voltage rises with negative-sequence current, trace the external system toward the utility connection. If terminal voltage remains closely balanced while current stays unbalanced, move next to phase impedance, internal generated voltage, and the measurement chain. Where a transformer lies between the generator and grid, record both sides; its winding connection and grounding determine which sequence components pass and how phase quantities map across it.

Check 3: Separate machine voltage from grid influence

Run an open-circuit voltage test only under the approved operating procedure, with the generator isolated from the utility and load. The test removes load-current voltage drops and grid forcing, exposing internal generated-voltage asymmetry.

  1. Open and verify the generator breaker and all other possible connections to the utility or load.
  2. Bring the generator to its normal test speed and excitation condition without closing onto the bus.
  3. Measure all three phase-to-phase voltages. Where the neutral is available and the grounding arrangement permits the test, also measure all three phase-to-neutral voltages.
  4. Use the same calibrated instrument and method on each phase. Record magnitude, frequency, and phasor angle rather than relying on panel-display rounding.
  5. Compare the measurements with the generator manufacturer's acceptance limits and earlier commissioning records.

Unbalanced open-circuit voltage sends the investigation into the replacement generator: stator winding connections, neutral connection, internal leads, voltage-sensing circuits, and winding condition. Balanced open-circuit voltage shifts the next check toward unequal series impedance, the interconnection, and CT circuits.

A voltage-regulation curve can support a cross-check. Use each measured phase current and the applicable generator regulation data to estimate the internal or no-load phase voltage required to produce the terminal reading. Large disagreement among the three calculated phase results points toward unequal impedance or incorrect current measurement. Apply only the manufacturer's curve for the installed generator and operating point; do not substitute a generic regulation percentage.

Check 4: Trace every phase conductor and connection

Work from the relay CT location toward the generator terminals and neutral, then toward the utility. Compare like-for-like connections on all three phases. The replacement followed a mechanical failure that damaged stator end turns, so installation details and every disturbed connection deserve direct inspection.

  1. Inspect generator terminals, bus joints, cable lugs, disconnect contacts, breaker stabs, and the neutral link for looseness, heating, contamination, or unequal contact area.
  2. Measure connection resistance using a method suitable for low-resistance joints. Compare phase results at the same temperature and with the same lead placement.
  3. Confirm that phase conductors have equivalent length, size, spacing, and routing.
  4. Look for a conductor routed around the opposite side of a steel support, enclosure member, or other substantial ferrous structure. A different magnetic path adds inductive impedance and can behave like an unintended reactor.
  5. Identify any transformer between the generator and utility. Record its actual winding connection, neutral treatment, tap positions, and phase mapping from the nameplate and drawings rather than assuming a connection.
  6. Record whether the generator neutral is solidly grounded, impedance grounded, or isolated. Grounding strongly affects zero-sequence current paths; negative-sequence current can still flow in a three-wire circuit, so neutral inspection does not replace phase-path checks.

If a joint or routing difference is found, repair it under the site's electrical safety procedure and repeat resistance and phasor measurements before returning the generator to parallel service.

Check 5: Prove the CT circuits end to end

Do not clear the measurement branch merely because two displays agree. Determine whether the relay and meter use independent CT cores and independent secondary wiring. If they share CTs, terminal blocks, or a common primary conductor arrangement, both can reproduce the same error.

  1. Compare CT ratios, accuracy classes, polarity marks, core assignments, and secondary burden for all phases.
  2. Inspect shorting blocks, test switches, terminal tightness, grounding points, and phase identification.
  3. Perform excitation or magnetization-curve testing on each CT and compare the curves for mismatch or damage.
  4. Apply secondary injection to verify relay channels, metering channels, scaling, phase assignment, and negative-sequence calculation.
  5. Where the outage and test equipment permit, perform primary injection through the complete CT chain. Primary injection tests CT ratio, polarity, wiring, input conversion, and displayed current together.

Swap only test connections under a controlled plan; uncontrolled phase swapping can hide a channel error or introduce a protection-polarity mistake. Restore every test switch and shorting link to its documented service position, then verify all three phase currents at zero load and during the next controlled loading step.

Restore production through controlled parallel testing

Return the unit to service only after correcting the identified path, machine, transformer, or measurement defect. Keep the approved negative-sequence protection active during the test.

  1. Capture baseline open-circuit voltages and confirm phase rotation before synchronizing.
  2. Synchronize under the site's operating procedure and begin at the lowest practical real- and reactive-power transfer.
  3. Record generator and bus voltage phasors, current phasors, I2, power factor or VAr output, excitation level, and stator temperatures from synchronized sources.
  4. Increase load in controlled steps. Stop if I2, excitation, or temperature approaches an approved limit, or if the current spread grows sharply with load.
  5. Compare the repaired phase currents with the 440 A, 450 A, and 505 A event values. Verify that relay and independent metering agree and that no negative-sequence element picks up.
  6. Retain the event files, test results, CT records, joint-resistance readings, and final phasor capture as the new baseline.

Get the unit running only after the resolving branch passes; then schedule permanent correction for any temporary conductor routing or connection repair. Do not normalize a protection setting around an unexplained current unbalance.

FAQ

Why does generator current become unbalanced when bus voltage looks balanced?

Current is driven by the small phasor-voltage difference across generator and interconnection impedance. A small negative-sequence voltage or unequal phase impedance can produce a much larger percentage current unbalance while the 480 V bus still appears nearly balanced.

Why does the AVR not correct one high generator phase current?

The AVR changes the common rotor field and therefore the aggregate reactive-power operating point. It does not independently modulate the three stator phases, so investigate negative-sequence voltage, unequal impedance, generator asymmetry, and CT errors.

Why can two meters show the same wrong generator current?

Separate analog-to-digital converters can still share CT cores, primary conductors, secondary terminal blocks, or wiring errors. Prove independence, then use excitation testing, secondary injection, and preferably complete-chain primary injection.

When should I stop testing generator current unbalance and escalate?

Stop if open-circuit voltage is unbalanced, CT or winding tests fail, current unbalance remains after path repairs, or controlled loading approaches the approved negative-sequence or temperature limit. Keep the generator out of parallel service and contact the generator, relay, transformer, or utility's official support channel with event files, phasors, CT test results, grounding details, and the one-line diagram.

Back to blog