MV Collector Protection When Inverter Fault Current Barely Exceeds Load

Do not set a renewable collector feeder's protection from a single assumed inverter fault-current multiple. Establish the minimum current actually seen by each relay, its duration and sequence content for every relevant operating state. If the minimum fault current overlaps permitted load current, increasing overcurrent sensitivity alone cannot provide both security and dependable fault clearance; another protection principle or operating restriction is needed.

1. Separate equipment duty from relay sensitivity

Maximum short-circuit current determines important switchgear and transformer duties. Minimum detectable fault quantities determine whether the protection can operate. These are different studies. A strong utility source may dominate a connected collector fault, while an islanded section, open incomer or weak-grid outage can leave mainly current-limited inverter contribution.

NREL's protection research explains that inverter response is programmable. SEL's field analysis also demonstrates why one generic current multiple is inadequate. Neither source supplies settings for this feeder. The PV interface guide gives broader connection context; this article concerns fault detectability at specific relay locations.

2. Obtain the missing manufacturer inputs

Request validated positive-, negative- and zero-sequence behaviour, phase-current limitation, reactive-current priority, voltage-dependent control, momentary cessation, recovery and firmware version. Distinguish grid-following and grid-forming modes. A phasor short-circuit model can screen cases, but controller transitions may require an electromagnetic-transient model and waveform-based relay evaluation.

Provide the collector topology, transformer vector groups, earthing arrangement, CT/VT data, cable impedances, grid equivalents and normal emergency loading. Zero-sequence current paths depend on transformer and grounding connections; do not infer them from the inverter's three-phase current rating. Retain the grounding-transformer guide as an interface reference.

3. Worked illustrative sensitivity gap

Consider an islanded 20 MVA collector block at 33 kV. Assume its total inverter contribution at the relevant relay is capped at 1.15 times rated current for the study interval. Assume permitted load reaches 1.10 times rated current and a proposed phase-overcurrent pickup is 1.20 times rated current. These are project hypotheses, not typical guaranteed inverter capabilities.

The proposed element will not pick up on this assumed fault contribution. Lowering pickup to 1.05 would place it below permitted load. The fault-to-load ratio is only 1.045, leaving little room for measurement uncertainty or transient loading. The calculation identifies a coordination problem; it does not calculate a valid replacement setting. In a grid-connected case the utility contribution could change these quantities entirely.

4. Choose a principle that survives the operating case

Differential protection still needs charging-current treatment and sensitivity checks. Directional elements are not automatically dependable when negative-sequence injection is suppressed or changes angle. Voltage supervision can improve logic but does not make every low-voltage event an internal fault. Select the architecture through the utility's protection philosophy and an explicit fallback plan.

5. Build a fault and outage matrix

Include three-phase, phase-to-phase and earth faults, near and remote locations, relevant fault resistance, minimum and maximum dispatch, charging and discharging BESS, controller modes and credible topology changes. Record which source reaches each CT; aggregate nameplate current is not necessarily relay current.

Use waveforms long enough to cover current limiting, cessation, ride-through and recovery. Check pickup, directional decision, trip time, reset and security on external faults. The transformer differential guide helps distinguish transformer-zone protection from collector feeder protection.

6. Procurement and commissioning evidence

Require model version control, the assumptions matrix, protection study, approved settings and executable replay cases tied to the supplied firmware. Obtain relay manufacturer review where a supervising element depends on unconventional current behaviour. Test communication loss and topology indication failures as well as successful trips.

At commissioning, retain secondary-injection records, end-to-end channel results, trip-circuit evidence and disturbance-recorder configuration. Factory simulation is not evidence that the installed CT polarity or intertrip wiring is correct. Conversely, a successful injection test does not validate an incorrect inverter fault model. IEC 60076-1 frames the transformer specification interface, not a universal collector relay setting.

7. Frequently asked questions

Do all inverters supply the same fault current? No. Controls, voltage, pre-fault state and equipment limits matter.

Is grid-forming mode sufficient to keep existing settings? No. Current limiting and sequence behaviour still need evaluation.

Can differential protection remove every uncertainty? No. It changes the detection principle, but CTs, communications, charging and backup must still be engineered.

8. References

• NREL: protection
• SEL: observed inverter fault response
• IEC: transformer interface scope