Numerical Relays: Wiring Skill Is Essential, Not Optional

The relay display shows a differential current that should not be present, the SCADA value disagrees with the injected current, or a protection element picks up at an unexpected level. Start with the displayed quantity, then trace its complete measurement path: screen binding, SCADA tag, communication mapping, relay data object, configured input, terminal, secondary wiring, and CT or VT. A numerical relay can expose the problem through records and vector diagrams, but it cannot correct crossed phases, reversed polarity, excessive burden, weak DC supply, or an incomplete trip circuit.

What is the screen telling you?

A displayed value is the result of several transformations. The relay samples an analog input, applies ratio and connection settings, calculates protection quantities, and publishes selected values to its local display, Web interface, or SCADA connection. The SCADA path adds a protocol mapping, driver, tag database, and screen binding. A correct number at one layer does not prove the next layer is correct.

For the documented differential case, additional measurements and vector analysis led back to incorrectly identified star-point CTs and a cyclic phase transposition in the secondary circuits. The calculated quantities were useful indicators; the repair point was outside the relay. The check passes when local raw measurements, calculated quantities, and phase labels agree with the known test condition.

Which settings connect the physical circuit to the calculation?

Before testing protection functions, record every setting that converts terminal signals into engineering quantities. A mistake here can make correct wiring look faulty or make crossed wiring appear plausible.

The tag can be right while the binding is wrong. Compare the local relay value with the protocol value before editing protection settings. If they match, move outward through the driver tag and screen binding. If they do not match, move inward through the relay point map, calculation, configured input, and terminals.

Modern relays may support event lists, GPS synchronization, TCP/IP communication, and IEC 61850. These features improve visibility and integration, but they remain separate from the electrical measurement path. The check passes when the same test quantity retains its magnitude, sign, phase identity, units, and timestamp from the relay to the operator screen.

How do you prove CT and VT secondary circuits first?

For one reported protection workload, an estimated 95% of faults involved CT or VT design and wiring, sizing, DC supply, control circuits, or trip circuits. Treat those circuits as primary commissioning targets, not as accessories to the numerical relay.

1. Review the drawings against the installed terminal designations. Mark CT cores, ratios, polarity marks, grounding point, phase sequence, and destination terminals.
2. Perform point-to-point continuity checks from each CT or VT terminal to the relay. Identify every transition through marshalling terminals, test switches, and cubicle wiring.
3. Prove polarity and phase identity using an approved injection method. Watch both numerical magnitude and phasor direction.
4. Calculate the complete secondary burden from conductor length, conductor area, terminals, test devices, and relay input. Compare it with the applicable CT or VT design data.
5. Measure voltage drop under the relevant test current. A continuity beep does not prove that the circuit performs correctly under load.
6. Perform insulation-resistance testing only with sensitive relay and communication electronics isolated as required by their instructions. Select the test voltage from the equipment documentation.
7. Inspect the DC supply and control wiring, then prove the trip path independently of the protection calculation.

A documented cubicle used a 30 m CT cable with 2.5 mm² conductors followed by 1 mm² internal cubicle wiring. Calling the circuit “only 5 A” misses the design question: resistance and burden depend on conductor material, cross-sectional area, total loop length, terminal resistance, and current. Calculate the installed loop rather than judging one cable segment in isolation.

The check passes when continuity, polarity, phase identity, burden, insulation condition, grounding, and loaded voltage drop all match the approved circuit design.

Why must a numerical relay be tested differently from an electromechanical relay?

An electromechanical overcurrent relay intended to trip at 1 A but operating at 1.2 A may require mechanical calibration through its adjustment components. Applying that habit to a numerical relay is wrong. Its pickup result depends on the injection source, input channel, configured nominal quantities, ratio scaling, test method, protection characteristic, output logic, and the manufacturer's stated tolerance.

1. Confirm whether the test plan uses primary or secondary values.
2. Verify the configured CT or VT ratio and nominal input reference.
3. Check the test source connections, phase angle, frequency, and measurement reference.
4. Observe element pickup separately from output contact operation and breaker operation.
5. Compare the measured pickup with the applicable device documentation. If it is outside specification, investigate calibration through the manufacturer's approved process; do not trim a setting merely to make one test point pass.

Numerical settings also need dimensional review. For a 200/5 CT, the ratio is 40:1. If 1456.7 A is a primary-referenced value, its ideal secondary equivalent is:

Isecondary = Iprimary × 5 / 200
Isecondary = 1456.7 × 5 / 200
Isecondary = 36.4175 A

That calculation does not determine whether the setting or injection is valid. Read the relay's setting reference, input capability, protection range, and test-set capability before applying it. A test that proves pickup around 4.99 A or 5.01 A still does not prove CT polarity, phase sequence, secondary voltage drop, or circuit insulation. The check passes when the tested quantity, units, reference side, and expected operating result are documented without hidden scaling.

How should the multifunction relay be divided into testable layers?

A multifunction numerical relay is effectively a protection cubicle concentrated into one device. The smaller enclosure does not reduce the number of functions that require testing. Divide commissioning into layers so a failed result points to a limited fault domain.

Testing the relay alone proves only part of the protection system. Testing every protection function without the external trip and measurement circuits also leaves critical failure modes untouched. The check passes when each layer has an independent result and the interfaces between layers have been exercised.

How do event records and vector diagrams support the diagnosis?

Numerical relays can produce more data than an operator can use at once. Filter it according to the question being answered. For unexpected differential current, retrieve phase magnitudes, angles, differential current, restraint current, channel mapping, and the event sequence. For an unwanted distance operation, review the measured trajectory and the configured resistive reach. For power-swing behavior, compare the captured trajectory with the configured swing-detection behavior.

Records can support setting corrections, including power-swing speed criteria and resistive reach for distance protection. Make a correction only after the record has been tied to verified instrument-transformer ratios, phase assignments, time references, and system conditions. Bad wiring produces detailed but misleading data.

Time synchronization improves correlation between relay events and SCADA records. A GPS-synchronized event list can help sequence operations across devices, while TCP/IP or IEC 61850 can transport data to higher-level systems. Neither communication method proves that the analog inputs are wired correctly.

The check passes when a known injection or switching event appears with the correct phase quantities, element transitions, output state, SCADA indication, and chronological order.

How do you perform the end-to-end commissioning test?

1. Freeze the approved drawings, protection settings, communication mapping, and test plan used for the test.
2. Verify CT and VT circuits point to point, including polarity, phase sequence, grounding, conductor transitions, and terminal identification.
3. Confirm relay input ratios, channel assignments, transformer compensation, protection settings, and output logic.
4. Inject known secondary quantities and compare the test source with raw relay measurements and vector diagrams.
5. Test protection pickup, restraint, timing, blocking, and output logic at documented operating and non-operating points.
6. Operate the physical output and trace the DC trip circuit through to the controlled device.
7. Confirm auxiliary contacts and relay inputs return the correct open, closed, healthy, or tripped state.
8. Compare local values and events with the communication object, SCADA driver tag, HMI binding, alarm, and event list.
9. Download the final settings, event records, and test results, then record the installed wiring corrections and final revision identifiers.

The final test must prove the whole chain, not separate demonstrations by different teams. Apply a known electrical condition, observe the correct relay measurement and protection decision, verify the intended output and controlled-device action, and confirm that the final state and event appear correctly on the operator screen.

FAQ