Home › Guides › Protection Relay Testing

Protection Relay Testing: Secondary and Primary Injection

A protection relay is the one device in a substation that is expected to do nothing for years and then work perfectly, in milliseconds, on the worst day of the transformer's life. There is no partial credit: a relay that trips a healthy transformer costs an outage, and a relay that fails to trip a faulted transformer can cost the transformer itself. Protection relay testing exists for that reason. It proves, before energisation and at intervals through life, that each relay measures correctly, that its settings are the ones the protection study intended, that its outputs reach the right breaker, and that the current transformers feeding it are wired with the right ratio and polarity.

This guide explains relay testing in plain language: secondary injection and primary injection, the main relay elements tested on transformer and feeder schemes, the test sets used, and the records that make a test worth anything afterwards. As with everything on this site, always work to your project specification, the manufacturer's instructions and your site safety rules; the general guidance here never overrides those documents. Relay testing is done by competent persons under permit, with tripping circuits made safe where required, and with current transformer secondaries never left open-circuited — an open CT secondary can develop a dangerous voltage. If you are new to transformer protection as a whole, start with our transformer protection guide and our power transformer testing and commissioning guide.

Why a factory-tested relay still needs site testing

It is a fair question, and it has a good answer: factory testing proves the relay, but site testing proves the installation. A relay can leave the factory in perfect condition and still fail on site because the wrong settings file was loaded, because a CT ratio was entered incorrectly, because two CT wires were swapped, because the trip output was wired to the wrong breaker, or because a test switch was left in the wrong position after someone else's work. Every one of those errors has caused real incidents, and every one of them is invisible until the relay is tested as installed.

Site testing therefore has two distinct purposes, and confusing them is a common mistake. The first is proving the relay element itself: given a known current and voltage, does it operate at the set threshold, in the set time, with the set characteristic? The second is proving the scheme: does a fault on the protected equipment actually result in the right breaker opening, the right alarms appearing, and nothing else tripping? Secondary injection answers the first question. Primary injection, together with trip circuit proving, answers the second. A commissioning programme that does only one of them is only half done.

Secondary injection: proving the relay

Secondary injection applies test currents and voltages directly to the relay's secondary terminals, simulating what the current and voltage transformers would deliver during a fault. A modern relay test set can generate three-phase currents and voltages with precise magnitude, phase angle and frequency, and can step them through a test sequence automatically while recording the relay's response. The relay is typically isolated from the live installation by its test switch or test plug, so that its outputs can be monitored without actually tripping a breaker — although a controlled trip test is usually performed as well, with the breaker made safe or the trip circuit monitored.

The tests performed depend on the relay elements enabled, but the pattern is consistent. Pickup is checked first: the current or voltage at which the element just begins to operate, compared with the setting. Then timing: at a multiple of pickup — commonly two, five or ten times — the operating time is measured and compared with the expected time from the relay's characteristic curve, whether definite-time or inverse-time. For directional elements, the operating boundary is checked by sweeping the phase angle and confirming the relay operates in the forward direction and restrains in the reverse. For differential elements, the test set simulates through-currents and fault currents and confirms the relay restrains for load and external faults and operates for internal faults, including the restraint characteristic at high through-currents.

Two details separate good secondary injection from box-ticking. The first is settings verification: before any test, the settings actually loaded in the relay are read back and compared, line by line, with the approved settings schedule. Testing a relay against the wrong settings proves nothing. The second is testing the complete output path at least once: not just confirming that the relay's internal logic operated, but that its output contact closed, that the trip circuit carried the signal, and that the intended breaker trip coil or lockout relay responded. A relay that "operates" in its own display but whose output wiring is broken protects nothing.

Primary injection: proving the whole chain

Primary injection goes one step further back: a large test current is passed through the primary side of the current transformers, so that the CT, its secondary wiring, the relay and the trip circuit are all proved together, exactly as they will behave during a real fault. The test current is typically injected by a high-current source connected across the primary conductors, with the protected equipment out of service and earthed. Because the current flows through the real CT, primary injection catches the errors secondary injection cannot: a CT wired with the wrong ratio, a CT polarity reversed, a CT secondary wire swapped between phases, or a CT that is saturated or damaged.

Polarity deserves emphasis, because it is the error primary injection exists to catch. Differential protection in particular depends entirely on CT polarity: if one CT is reversed, the relay sees a large differential current during normal load and will trip a healthy transformer, or, depending on the scheme, will fail to see a real fault. Secondary injection cannot reveal this, because the test set feeds the relay directly and bypasses the CTs. Primary injection, by contrast, reveals it immediately: with current flowing in the real primary circuit, the relay's measured currents can be checked for correct magnitude and phase relationship, and a reversed CT shows up as a current pointing the wrong way. This is why primary injection is not an optional extra on differential schemes — it is the test that proves the scheme is wired to protect rather than to nuisance-trip.

Primary injection is heavier work than secondary injection: it needs a high-current source, careful earthing arrangements, and coordination with the rest of the outage programme, since the primary conductors under test are part of the equipment being commissioned. It is also, by necessity, done at currents well below full fault level — enough to prove ratio and polarity reliably, not enough to stress the equipment. The results are recorded as measured secondary currents and phase angles for a stated primary current, so that they can be compared with the expected values from the CT ratios. A result that is wrong by the ratio of two CTs, or that shows one phase reversed, is not a borderline result to be argued about; it is a wiring error to be found and fixed before energisation.

The main relay elements, and what each test proves

Overcurrent and earth-fault elements are the workhorses of most schemes, and their tests are straightforward: pickup, timing at several multiples, and, for earth fault, confirmation that the element responds to the correct residual current and not to balanced load. Instantaneous elements are checked for pickup and for the absence of unwanted delay. Where overcurrent is directional, the directional boundary is proved as described above.

Differential elements need the most careful testing. Beyond the basic pickup and restraint checks, the test set should prove the second-harmonic restraint that stops the relay tripping on transformer magnetising inrush at energisation — a feature covered in our transformer protection guide. If the restraint is set wrong, the relay either trips on every energisation or fails to restrain when it should, and both failure modes have been seen in service. Restricted earth fault elements are tested similarly, with attention to the stabilising resistor and the operating threshold.

Other elements tested as applicable include under- and overvoltage, frequency, breaker failure, and Buchholz and temperature trip circuits — the last of these being proved end-to-end from the device contact through to the breaker, since a Buchholz trip that never reaches the breaker is a protection gap, not a protection scheme. Our Buchholz relay guide covers the device itself; relay testing covers the circuit it trips through. In every case, the principle is the same: test the element, test the output, and test the complete path at least once.

Test sets, records and acceptance

Modern relay test sets are computer-controlled instruments that generate precise currents and voltages, automate test sequences, and record results against expected values. Older sets, and the manual methods still used for simple relays, achieve the same ends more slowly. The instrument matters less than the method: a test set is only as good as its calibration, and results taken with an out-of-calibration set are not defensible. Calibration certificates, with dates, should be part of the test record package, and any result that is out of tolerance should be re-measured with a different set before the relay is condemned.

Acceptance criteria come from three places: the manufacturer's stated accuracy, the project specification's tolerances, and the protection study's intent. A pickup within a few percent of setting and a timing within the manufacturer's tolerance band is normally accepted; a result outside it is investigated, not filed. The investigation order is always the same: check the test setup, check the settings, check the wiring, and only then suspect the relay. Most out-of-tolerance results turn out to be one of the first three.

Records are what turn a morning of testing into evidence. A complete relay test record shows the relay identification, the settings read back from the relay, the test set and its calibration date, the test currents and voltages applied, the measured pickup and timing values, the expected values, the pass or fail verdict for each element, the trip circuit proved, and the name of the tester and the witness. Those records are what the next engineer, the asset owner and, in the worst case, an incident investigation will rely on. A relay test with no record is, for practical purposes, a relay test that never happened.

Frequently asked questions

Related guides: power transformer testing and commissioning guide, winding resistance test, loop checks explained, substation commissioning guide.