Transformer Energization: Inrush Current vs. Internal Faults How 87T Relays Tell the Difference

November 05, 2026 | Blog

Part A — Transformer Energization at a Glance

A breaker closes. A large power transformer that has been out of service for maintenance, or a new main power transformer at a solar or battery plant, is energized from the 230 kV bus. Within a quarter cycle, the differential relay sees several times rated current flowing into the transformer on one side and almost nothing coming out of the other. By every simple measure, that looks like an internal fault.

It usually is not. It is magnetizing inrush current: a large, distorted, slowly decaying current that the transformer core draws while its magnetic flux settles into steady state. It is real primary current, not a measurement error, and it flows only on the energized side. To a differential relay (device 87T), it appears entirely as differential current.

The engineering challenge is simple to state and hard to solve: block or restrain for inrush every time, but still trip in a cycle or two for a genuine internal fault — including a fault that is present at the moment of energization. Getting it wrong in one direction trips a healthy transformer and creates a reportable misoperation. Getting it wrong in the other direction leaves a faulted transformer energized while the fault grows.

This guide explains, in US engineering terms, what drives inrush, how large it gets, why the classic second-harmonic rule is no longer sufficient on its own for modern low-loss transformers, and which discrimination methods modern 87T relays use. It also covers controlled (point-on-wave) switching, commissioning practice and the NERC compliance consequences of getting it wrong. Keentel Engineering's own simulation of a 100 MVA, 230/69 kV transformer makes each point concrete.

Keentel's companion article on transformer differential security covers 87T fundamentals, compensation and CT saturation during external faults in depth. This article focuses on energization.

The transformer used in this guide

What Keentel's analysis shows

Figure 1. Keentel's simulation of the 100 MVA, 230/69 kV transformer. Top: line currents into the 230 kV delta winding when energized with high residual flux on a modern high-flux-density core. Middle: an internal fault current, shown as a separate case for comparison. Bottom: second-harmonic content of each differential current, calculated with a one-cycle window, against a 15% blocking threshold.

Engineering review of common statements about inrush and 87T

Five lessons

• Inrush is real current, not an error. The relay is measuring correctly; it must recognize what it is measuring.
• Second harmonic is a symptom, not a law. The classic idea that inrush always carries at least about 15–20% second harmonic dates from older cores. Settings must be checked against the actual transformer and, ideally, recorded energizations.
• Cross-blocking buys security and costs a little dependability. If any phase blocks all phases, a fault during energization may be delayed until the inrush decays — the unrestrained element and other logic must cover that case.
• The unrestrained element must clear the worst inrush — by the quantity it actually measures. In Keentel's example the fundamental is 4.5 pu but the instantaneous peak is 8.8 pu.
• The best inrush is the one you never create. Point-on-wave closing, energizing from the stronger side and sequencing parallel transformers reduce stress on the transformer, the relay and the grid.
  

Part B — Technical Guide: Inrush Physics and How 87T Relays Tell It Apart From Faults

1. Where the 87T relay stands at the moment of energization

Transformer differential protection compares the currents entering and leaving the transformer after correcting for ratio, the delta-wye phase shift and zero-sequence current. A modern numerical relay computes an operating current and a restraint current for each phase:

and trips when the operating current exceeds both a minimum pickup and a slope times the restraint current. Keentel's companion article on transformer differential security explains this characteristic, compensation and CT saturation during external faults.

At energization, the arithmetic is unforgiving. Only one winding carries current. If the 230 kV breaker closes with the 69 kV breaker open, the restraint current (by the average definition above) is half the operating current. The operating point therefore sits on a line with a slope of 200% — far above any percentage-restraint slope. A percentage-differential element alone cannot be secure against inrush. Some additional quantity — harmonic content, waveshape, DC content or flux — must tell the relay that this is energization, not a fault.

2. The physics of magnetizing inrush

2.1 Flux is the integral of voltage

Faraday's law links the flux linkage of a winding to the voltage applied across it. Ignoring the small resistive drop:

where λr is the residual (remanent) flux left in the core when it was last de-energized and t0 is the closing instant. With a sinusoidal voltage of peak value Vm, the steady-state flux is a sinusoid of peak value Vm/ω that lags the voltage by 90 degrees. Expressed per unit of rated peak flux, steady-state flux swings between −1 and +1 pu.



The core does not start in steady state. At the closing instant, the flux equals the residual flux, while the steady-state solution would demand some other value. The difference appears as a DC flux offset:

in per unit, where θ is the angle of the voltage waveform at closing (with the voltage written as a cosine). The DC term λr − sin θ decays only through winding and source resistance, so it persists for many cycles.

2.2 The worst case: closing at voltage zero

If the breaker closes when the voltage passes through zero heading positive (θ = −90 degrees), the flux starts at λr and rises by 2 pu over the next half cycle:

With residual flux of 0.8 pu in the same direction, the peak demand is 2.8 pu — almost three times rated flux. No practical core can carry that without saturating.

2.3 Saturation and the air-core reactance

Grain-oriented silicon steel saturates at roughly 1.9–2.0 tesla. Large power transformers are designed to operate at about 1.5–1.75 tesla at rated voltage, so the saturation flux is roughly 1.15–1.35 pu of rated peak flux. Modern low-loss, high-flux-density designs sit toward the lower end of that range — closer to saturation — which, as Section 5 shows, matters for protection.

Below saturation, the magnetizing inductance is enormous and magnetizing current is a fraction of a percent of rated current. Once the core saturates, the flux the core cannot carry must pass through air, oil and winding space, and the winding behaves like an air-core inductor. The air-core reactance of the energized winding is typically a few tenths of a per unit — larger than the leakage reactance between windings but tiny compared with the unsaturated magnetizing reactance. Current rises steeply until the flux falls back below saturation, producing the characteristic unidirectional pulse each cycle.

2.4 A practical peak-inrush formula

A widely used estimate of the first peak, for a single-phase or per-phase view, combines these ideas:

where V is the rms applied voltage per winding, Xac is the air-core reactance of the energized winding plus the source reactance, Br the residual flux density, Bm the rated peak flux density and Bs the saturation flux density. Expressed per unit of rated rms current, √2·V/Xac becomes √2/Xac,pu.

2.5 Worked example 1 — peak inrush of a 100 MVA, 230/69 kV transformer

Rated current and base impedance at 230 kV:

Take an air-core reactance, including the source, of 0.25 pu (about 132 Ω). For a modern high-flux-density core with Bs/Bm = 1.15 and residual flux of 0.8 pu:

Keentel's time-domain simulation of the same transformer, energized from the delta-connected 230 kV side with a piecewise-linear saturation curve, gave a worst-phase peak of 8.8 × rated rms current — within about 5% of the formula. Typical published ranges for large power transformers are roughly 6 to 12 times rated current for the first peak; small distribution transformers can reach 20 to 30 times their (much smaller) rated current because their air-core reactance in per unit is lower relative to their rating.

The formula shows why inrush is so variable from one energization to the next: residual flux and closing angle are effectively random unless controlled, so the same transformer can draw nearly zero inrush one day and nine times rated current the next.

3. What shapes the inrush waveform

4. How fast inrush decays



The DC flux offset decays because each unidirectional current pulse produces a voltage drop across the circuit resistance, which slowly pulls the average flux back toward zero. While the core is saturated, the relevant time constant is roughly:

but the core is saturated for only part of each cycle, so the effective decay is slower than this expression suggests and changes as the inrush shrinks. In practice:

• Small transformers decay within a few cycles to a fraction of a second.
• Large power transformers fed from strong, low-resistance transmission sources can take several seconds, and the tail of the inrush can persist for tens of seconds.
• Sympathetic inrush (Section 6.1) can last longer still, because the DC current circulates between transformers through a low-resistance path.

4.1 Worked example 2 — decay in Keentel's model



Keentel's simulation used a moderately damped circuit to keep the figure readable. The worst-phase peak fell as follows:

A real 100 MVA transmission transformer on a stiff 230 kV bus would typically decay more slowly. Two practical consequences follow. First, any blocking or restraint logic must remain effective for as long as the inrush lasts, not just for the first few cycles. Second, the second-harmonic ratio rises as inrush decays (bottom panel of Figure 1), so the hardest moment for harmonic-based security is the first few cycles, when the current is also largest.

5. Three-phase inrush and the harmonic content the relay sees

5.1 Why the three phases differ

In a three-phase transformer, the three poles of an uncontrolled breaker close at three different points on the voltage wave, and the three core limbs start with three different residual fluxes. The phases therefore saturate by different amounts and at different times. In the example, residual flux of +0.8, −0.8 and 0 pu on the three winding limbs is a typical pattern left behind by de-energization.

The connection then mixes the winding currents before the relay sees them:

• On a delta-connected winding, each line current is the difference of two winding currents. A line can carry two saturation pulses of opposite polarity in each cycle, which fills in the zero-current gaps and changes the harmonic content.
• On the grounded-wye side, the relay removes zero-sequence current and applies the vector-group compensation matrix, which also subtracts phase currents.

The result is that the differential current in one phase can look much more like a symmetrical, fault-like waveform than the classic unidirectional inrush pulse — with less second harmonic.

5.2 Worked example 3 — harmonic content of the simulated inrush

Keentel computed a one-cycle Fourier transform of each differential current (the 230 kV line currents, since the 69 kV winding carries no current) for two cores, energized at the same closing angle.

With the conventional core, every phase is comfortably above a 15% threshold. With the modern high-flux-density core, phase B — the phase carrying the largest current — has only about 10% second harmonic in the first cycle and stays below 15% for about four cycles. The fourth harmonic adds almost nothing in this case, so a harmonic-sum method would not rescue phase B on its own.

The phase with the most inrush current is not necessarily the phase with the most second harmonic. Often it is the opposite.

5.3 Why second harmonic exists at all

A waveform with half-wave symmetry — where the negative half cycle is a mirror image of the positive half — contains only odd harmonics. Fault current, load current and the excitation current of a transformer in steady state are all close to half-wave symmetric. Inrush is not: the DC flux offset makes the core saturate on one half cycle only, producing unidirectional pulses separated by gaps. That asymmetry is what creates even harmonics, particularly the second.

The second-harmonic content of a train of unidirectional pulses depends on the pulse width. Narrow pulses with long gaps (light saturation) have high second harmonic, often 30–60%. Wide pulses with short gaps (deep saturation, typical of high-flux cores with high residual flux) have lower second harmonic. Older literature and many legacy settings assumed inrush would always carry at least about 15–20% second harmonic. Modern low-loss transformers and certain delta connections can produce less than 10% in one or more phases, and published field records show values as low as about 7%.

5.4 Internal faults are not perfectly sinusoidal either

An internal fault draws current through the relatively linear leakage and system impedance, so the waveform is close to a sinusoid. But it carries a decaying DC offset whose size depends on the fault inception angle and the X/R ratio. In Keentel's example (X/R = 20), a one-cycle Fourier filter reports about 4% second harmonic in the first cycle, falling toward zero. If the CTs saturate during a heavy internal fault, the secondary waveform becomes asymmetrical and can carry significant second harmonic — which is exactly why harmonic blocking can delay tripping for severe internal faults and why the unrestrained element exists.

6. Related energization phenomena

6.1 Sympathetic inrush

When a transformer is energized in parallel with one already in service, the DC component of the incoming transformer's inrush flows through the shared source resistance. That produces a DC voltage drop across the in-service transformer, which slowly drives its core flux in the opposite direction until it also saturates on alternate half cycles. The two transformers then exchange a slowly decaying, low-harmonic current that can last many seconds. The in-service transformer's 87T sees a differential current with low second harmonic, and its CTs see prolonged DC. Sympathetic inrush has caused misoperations at plants with two or more main power transformers on a common bus — a configuration common at large US solar, wind and battery projects and at data center substations.

6.2 Recovery inrush

When a nearby external fault is cleared, the voltage on an in-service transformer recovers abruptly. The flux must re-establish itself, and an inrush-like current flows, usually smaller than energization inrush because there is no residual-flux mismatch of the same size. Restraint logic must cover recovery inrush, which arrives at the worst possible moment — right after a through fault that may have left CTs with remanence.

6.3 Overexcitation is a different problem

Overexcitation occurs when voltage is high or frequency is low, so the volts-per-hertz ratio drives the core toward saturation symmetrically on both half cycles. The resulting excitation current is half-wave symmetric, rich in fifth harmonic and nearly free of second harmonic. IEEE C57.12.00 requires transformers to operate at 105% of rated secondary voltage at rated load and 110% at no load; beyond these limits, core heating rises quickly. Differential relays commonly block the 87R element when fifth harmonic exceeds a setting (often around 30–35%), and a separate volts-per-hertz element (device 24) provides the actual overexcitation protection, following the guidance of IEEE C37.91 and the transformer manufacturer's V/Hz withstand curve.

6.4 CT performance during long inrush tails

A CT is rated for transient performance assuming a fault DC offset with a time constant of tens of milliseconds. Inrush DC can persist for seconds. The flux a CT must develop to reproduce a slowly decaying DC component grows with the product of the DC current and its time constant.

6.5 Worked example 4 — can the 230 kV CT reproduce the inrush tail?

Assume the inrush carries an average (DC) component of about 3 pu of rated current — 750 A primary — decaying with an effective time constant of 0.5 s, measured by a 400:5 CT with a total secondary burden of 2 Ω. The equivalent X/R of the DC component is ωτ = 377 × 0.5 ≈ 188. Using the same logic as the IEEE C37.110 transient-factor check:

That is several times the capability of even a C800 CT. In practice the inrush tail decays and the CT does not see a pure DC source, but the result shows why long inrush tails — especially sympathetic inrush — can drive CTs into partial saturation, distorting the measured waveform and its harmonic content. EMT studies with realistic CT models are the right tool when this is a concern.

7. How modern 87T relays discriminate inrush from internal faults

7.1 Second-harmonic restraint versus blocking

Two classic approaches use the same measurement in different ways:

• Harmonic restraint adds a multiple of the harmonic current to the restraint quantity, so the trip condition becomes, for example:
  

• Harmonic blocking compares the harmonic ratio directly with a threshold and blocks the restrained element while it is exceeded:
  

Typical second-harmonic settings are 12–20%, with 15% a common starting value; some utilities set lower values for modern transformers after reviewing energization records. Lower thresholds increase security but also increase the chance that a fault with CT saturation is delayed.

7.2 Independent, cross-blocking and averaging schemes

7.3 Worked example 5 — does the relay trip?



Settings: minimum pickup 0.3 pu of tap; second-harmonic threshold 15%. The first-cycle differential currents from Worked example 3 (high-flux core):

Without cross-blocking, phase B's restrained element operates within the first cycle or two — a misoperation on a healthy transformer. With cross-blocking, all three phases stay blocked until phase B's own second harmonic rises above 15% at about 71 ms, by which time phases A and C are well above threshold. Lowering the threshold to 10% would also have secured this case, but at the cost of security margin elsewhere and dependability for faults with CT saturation. The right answer depends on the transformer, which is why recorded energizations are so valuable.

7.4 Fourth harmonic and harmonic sum

Even harmonics other than the second are also signatures of waveform asymmetry. Some relays add the fourth harmonic to the second, either as a phasor or arithmetic sum, or as a separate restraint term. In Keentel's example the fourth harmonic was only 1–2%, so adding it would not change the outcome. In other transformers — particularly with wider saturation pulses — the fourth harmonic contributes more, and harmonic sum restraint closes part of the gap left by a low second harmonic.

7.5 Waveshape recognition and dead-angle detection

Inrush current stays near zero for part of each cycle while the core is out of saturation; fault current does not. Waveshape-based relays measure the duration of these low-current intervals — the dead angle or wave gap — and declare inrush when the gap exceeds a threshold. A quarter cycle (90 degrees) is a common reference, although thresholds are manufacturer-specific and some designs use smaller values to cover delta-connected windings, where gaps are partly filled by the second winding's pulse. Waveshape methods respond within about a cycle and do not depend on the transformer's harmonic content, but they can be affected by CT saturation, which itself creates gaps, and by the delta-connection effect.

7.6 DC and flux-based methods

Inrush carries a large, slowly decaying DC component; most internal fault currents carry a DC offset that decays in tens of milliseconds. Some relays use the DC content of the differential current as an additional restraint. More advanced methods estimate the transformer's magnetizing inductance or flux–current trajectory from measured voltage and current: an energizing transformer shows a large, characteristic change in apparent magnetizing inductance between saturated and unsaturated parts of each cycle, while a faulted winding shows a low, nearly constant inductance. These methods need voltage inputs and good models, and are offered in some modern relays.

7.7 The unrestrained element (87U)

The unrestrained differential element trips without any restraint or harmonic supervision when the differential current is so large that it can only be an internal fault. It provides fast, dependable tripping for heavy internal faults, including those where CT saturation creates harmonics that would otherwise block the restrained element. It must be set above:

• the maximum inrush current, measured the way the element measures it;
• the maximum false differential current from CT saturation during external faults (see Keentel's companion article on 87T security).

7.8 Worked example 6 — setting the 87U element



In Keentel's simulation, the largest one-cycle fundamental differential current during inrush was 4.5 pu of tap, while the largest instantaneous peak was 8.8 × rated rms current.

Typical 87U settings are 8–12 times tap. The example shows why the setting must be checked against the measuring principle of the specific relay and against the worst energization — including sympathetic inrush and inrush on the lower-impedance winding — not taken from a typical value alone.

7.9 Negative-sequence differential and restricted earth fault

Turn-to-turn faults involve a small number of shorted turns. The terminal current changes very little, so the phase differential elements may not detect them — and if the transformer is energized onto a turn fault, harmonic blocking may hold them off anyway. A negative-sequence differential element (87Q) compares the negative-sequence currents of the windings and is highly sensitive to unbalanced internal faults. Because inrush also contains negative-sequence current, 87Q elements are usually supervised or briefly delayed during energization, with logic specific to each relay.

Restricted earth fault protection (87N or REF) compares the transformer neutral CT current with the residual current from the phase CTs of a grounded-wye winding. It provides sensitive ground-fault coverage near the neutral, where phase differential elements are least sensitive. Inrush on a grounded-wye winding can contain zero-sequence current, and phase CT saturation can create false residual current, so REF elements are normally supervised by directional logic or set with security margins verified against energization records.

7.10 Summary of discrimination methods

8. When a fault is present at energization

The hardest case for any discrimination scheme is energizing a transformer that already has an internal fault — for example, after a repair, after a winding was damaged by a through fault, or when a grounding device or temporary connection was left on. The differential current is then the sum of inrush and fault current.

• On a faulted phase with heavy fault current, the fundamental component dominates, the harmonic ratio falls, and the restrained element can operate within a cycle or two — unless cross-blocking from a healthy phase holds it.
• With cross-blocking, tripping may wait until the healthy phases' inrush has decayed below their own thresholds, which can take several cycles or longer.
• The unrestrained element covers severe faults immediately. For lower-level faults, relays offer time-limited cross-blocking, independent blocking after a short time, or 87Q and REF elements that respond to the fault's unbalance.

The setting decision is therefore a deliberate trade-off between security for every energization and dependability for the rare energization onto a fault. Keentel recommends documenting that trade-off in the settings basis, because it is exactly the question a NERC misoperation review will ask.

9. Avoiding inrush: controlled switching

9.1 The principle

Inrush is caused by a mismatch between the core's residual flux and the flux the applied voltage demands at the closing instant. If the breaker closes at the instant when the prospective steady-state flux equals the residual flux, no DC offset is created and no inrush flows. For a single winding with no residual flux, that instant is the voltage peak, where prospective flux passes through zero.

With residual flux λr (per unit), the optimal closing angle satisfies:

on the portion of the wave where flux is rising in the same direction as the residual flux.



9.2 Worked example 7 — closing angle for 0.8 pu residual flux

At that instant the voltage is cos 53.1° = 0.6 pu and rising flux matches the residual flux. A closing-time error of 1 ms equals 21.6 degrees at 60 Hz. Closing 1 ms late creates a flux offset of sin(74.7°) − 0.8 = 0.165 pu, so peak flux reaches about 1.17 pu — just above the 1.15 pu saturation level of the high-flux core in the example. Controlled switching therefore needs breaker closing-time accuracy of well under a millisecond, which means compensating for temperature, control voltage, idle time and mechanical wear.

9.3 Three-phase strategies

Residual-flux-aware controllers estimate residual flux by integrating the winding voltage during de-energization (using capacitor-voltage or other suitable voltage transformers) and then choose the closing instants accordingly. Utilities commonly report inrush reductions from several per unit to a small fraction of rated current when controlled switching is properly commissioned.

9.4 Other mitigation options

• Pre-insertion resistors in the breaker damp the inrush and the voltage sag, at the cost of breaker complexity and maintenance.
• Energize from the higher-impedance side: the outer winding, usually the high-voltage winding, has a higher air-core reactance, so energizing from it limits peak inrush.
• Energize from the stronger bus when voltage sag is the concern, and from the weaker side when inrush magnitude is the concern — the two goals pull in opposite directions and must be balanced in the study.
• Sequence parallel transformers and, where possible, de-energize or isolate an in-service transformer's load-side connection to limit sympathetic inrush.
• Reduce residual flux: some operating procedures de-energize transformers in a way that leaves lower remanence, though this is rarely controllable on transmission transformers.

10. Grid impacts: voltage sag during energization

Inrush draws mostly reactive power at the fundamental frequency, so energizing a large transformer depresses the bus voltage. A quick estimate uses the inrush apparent power and the short-circuit capacity of the bus:

10.1 Worked example 8 — sag from energizing the 100 MVA transformer

The first-cycle fundamental inrush of about 3 pu corresponds to roughly 300 MVA:

Utilities typically limit rapid voltage changes from routine switching to a few percent — often in the range of about 3 to 5% depending on the voltage level and how often the event occurs, with specific values set in the transmission owner's interconnection requirements and in IEEE 1453 flicker guidance. Inverter-based plants with several main power transformers, and data center campuses with multiple large transformers, may need controlled switching, pre-insertion resistors or sequenced energization to stay within those limits. The same sag can disturb nearby sensitive loads and, if severe, interact with ride-through settings of nearby inverter-based resources.

10.2 Inverter-based plants and large-load campuses

US solar, wind and battery projects typically connect through one or more 230/34.5 kV or 345/34.5 kV main power transformers, with dozens to hundreds of 34.5 kV collector pad-mounted transformers downstream. Energization planning at these sites should consider:

• Main power transformer energization from the transmission system, with voltage sag and 87T security reviewed in the interconnection studies;
• Sympathetic inrush when a second main power transformer is energized beside one already in service;
• Collector feeder energization, where closing a 34.5 kV feeder breaker energizes many pad-mounted transformers at once, producing a large combined inrush with long decay that can trip feeder overcurrent elements or the main transformer's differential if not coordinated;
• Harmonic resonance: inrush harmonics can excite resonances formed by cable capacitance and the source, producing temporary overvoltages — a topic for EMT study.

11. Studying inrush: EMT simulation

Steady-state short-circuit programs cannot represent inrush. Electromagnetic transient (EMT) simulation can, provided the model includes:

• a nonlinear magnetizing characteristic for each core limb, ideally with hysteresis so residual flux can be represented and changed by de-energization;
• the correct air-core inductance of each winding (from the manufacturer or estimated from design data);
• the core type (three-limb, five-limb, shell) and winding connections;
• the source impedance, including other transformers that can produce sympathetic inrush;
• breaker pole-closing times, statistically varied over many energizations;
• CT models with saturation, so the relay sees realistic secondary currents;
• the relay's actual filtering and logic, or a recorded COMTRADE playback through the relay itself.

A statistical study of hundreds of energizations with random closing angles and residual fluxes gives the distribution of peak inrush, minimum second harmonic per phase and worst-case voltage sag — far more useful for setting decisions than a single "worst case."

12. Commissioning and settings review

Energization is the one time the protection engineer can observe the transformer's inrush under real conditions. Keentel recommends:

1. Review settings before first energization: compensation, slopes, harmonic thresholds, cross-blocking mode, 87U level and its measuring principle, fifth-harmonic blocking and the 24 element.

2. Trigger oscillography on breaker close and record differential currents, harmonic quantities and element states in COMTRADE format on every energization.

3. Energize more than once where practical — at least two or three times — because residual flux and closing angle change each time.

4. Analyze each record: peak inrush per phase, minimum second harmonic per phase, duration below the threshold, 87U margin, and whether any element came close to operating.

5. Check CT behavior during the inrush tail, especially at sites with parallel transformers.

6. Update the settings basis with the measured minimum harmonic ratio and the margin to the threshold, so future reviews start from evidence.

7. Repeat the review after major changes — a replacement transformer, a new parallel unit, a relay firmware change or a change in source strength.

13. NERC compliance: PRC-004 and PRC-005

For Bulk Electric System transformers, an 87T operation on energization of a healthy transformer is an unnecessary trip and therefore a protection system misoperation under NERC Reliability Standard PRC-004. The owner must identify the misoperation, determine its cause, and develop and implement a corrective action plan — for example, enabling cross-blocking, adjusting the second-harmonic threshold, revising the 87U setting or adopting controlled switching — and evaluate whether the same cause applies to other transformers with similar relays and settings. A failure to trip for a fault during energization is also a misoperation.

PRC-005-6 requires maintenance and testing of the protective relays, CTs, DC supply and trip circuits that make up the transformer protection system, at intervals set by the standard or by a documented performance-based program. Commissioning energization records and settings files are valuable evidence for both standards. Keentel's companion articles on sudden pressure relaying and on 87T security cover the PRC-005 aspects in more detail.

14. A practical checklist for 87T inrush security

• Obtain the transformer's saturation characteristic, air-core reactance and core type from the manufacturer.
• Estimate peak inrush for the worst residual flux and closing angle, on each winding that may be energized.
• Run an EMT study for large or critical transformers, including sympathetic inrush with parallel units.
• Choose a second-harmonic threshold and blocking mode (independent, cross, two-of-three, harmonic sum) based on the expected minimum harmonic ratio, not a generic value.
• Check the 87U setting against inrush measured the way the element measures it.
• Decide how a fault during energization will be cleared, and document the trade-off.
• Set fifth-harmonic blocking and the volts-per-hertz element for overexcitation.
• Specify CTs with transient margin for both external faults and inrush tails.
• Consider controlled switching or pre-insertion resistors where inrush, voltage sag or repeated misoperations justify them.
• Record every energization, analyze it and update the settings basis.
  

Part C — Three Anonymous Case Studies

These composite case studies are based on the kinds of problems Keentel Engineering sees on US projects. Names, locations and details have been changed or combined, and figures are rounded.

Case study 1 — A replacement transformer that tripped on energization

Situation. A Midwest utility replaced a failed 115/13.8 kV, 40 MVA distribution substation transformer with a new low-loss unit. The existing numerical differential relay, with its long-standing settings, was reused. On the first energization the 87T operated. The transformer passed its post-trip tests, was energized again and tripped again. The event was reported as a misoperation under NERC PRC-004 because the substation was part of a Bulk Electric System network path.

Analysis. Keentel reviewed the relay's event records. Peak inrush was about 8 times rated current. Two phases showed 22–30% second harmonic, but the third phase — the one carrying the most current — showed only 11–13% for the first five cycles. The relay was set for independent per-phase blocking with a 20% threshold, a setting inherited from the original transformer. The replacement core operated at a higher flux density, and its inrush simply did not carry the second harmonic the old settings assumed.

Engineered solution. Keentel enabled cross-blocking with a 15% threshold, added fourth-harmonic restraint, confirmed the 87U setting of 10 times tap against the recorded fundamental and peak quantities, and verified dependability by replaying simulated internal faults — including a fault during energization — through the revised settings. The settings basis was updated to record the measured minimum second-harmonic ratio.

Outcome. Five subsequent energizations were recorded without operation, with a minimum second-harmonic margin of about 4 percentage points on the low phase. The utility adopted a rule to review harmonic settings whenever a transformer is replaced, even when the relay is not.

Case study 2 — Sympathetic inrush at a solar and storage plant

Situation. A 300 MW solar and battery plant in the Southwest connects to a 345 kV substation through two 345/34.5 kV main power transformers on a common 345 kV bus section. During commissioning, the first transformer was energized and loaded. When the second transformer was energized several weeks later, the first transformer's differential relay tripped about 1.5 seconds after the breaker closed.

Analysis. Oscillography showed classic sympathetic inrush. The in-service transformer's differential current grew slowly over the first second, with second harmonic below 10% and a large DC component, while its 34.5 kV CTs showed signs of partial saturation from the long DC tail. The relay's time-limited cross-blocking had already expired, because it was triggered only by energization of the relay's own transformer.

Engineered solution. Keentel built an EMT model of both transformers, the 345 kV source and the CTs, and reproduced the event. The solution combined three measures: waveshape-based inrush detection enabled alongside harmonic blocking; cross-blocking without a short time limit, backed by a negative-sequence differential element for internal-fault dependability; and an energization procedure in which the in-service transformer is lightly loaded and the incoming transformer is energized with controlled closing on its independent-pole 345 kV breaker.

Outcome. Subsequent energizations produced peak inrush below 2 times rated current and no relay operation. Voltage sag at the 345 kV bus dropped from about 4% to under 1%, well inside the transmission owner's limit.

Case study 3 — A data center campus with a voltage-sag problem

Situation. A data center developer in the Mid-Atlantic built an owner substation with three 230/34.5 kV, 120 MVA transformers fed from a 230 kV line with relatively modest short-circuit strength. The interconnection study flagged that energizing any one transformer could depress the 230 kV voltage by more than the transmission owner's rapid-voltage-change limit, and the owner's own 34.5 kV loads were sensitive to sags.

Analysis. Keentel's statistical EMT study of 400 random energizations showed a worst-case first-cycle inrush of about 9 times rated current and a 230 kV sag of up to about 11%. It also showed that energizing a second transformer beside one already in service could produce sympathetic inrush lasting several seconds.

Engineered solution. Keentel specified independent-pole 230 kV breakers with residual-flux-aware controlled switching, a commissioning procedure to verify closing-time accuracy, and 87T settings with cross-blocking and harmonic sum restraint, validated against the study's lowest-harmonic cases. An energization sequence was written into the operating procedures, with alarms if the controller was out of service.

Outcome. Measured inrush during commissioning stayed below about 1.5 times rated current, the voltage sag was under 2%, and the transmission owner accepted the energization plan. The project avoided an expensive pre-insertion resistor retrofit.

Part D — 20-Question Technical FAQ

Keentel Services, References and Contact

How Keentel Engineering Helps

Keentel Engineering supports utilities, independent power producers, data center developers and industrial owners with transformer protection from specification through commissioning:

• Transformer protection settings: 87T, 87U, 87Q, REF, 24 volts-per-hertz and backup overcurrent settings, with harmonic blocking modes chosen from evidence rather than defaults.
• Energization and inrush studies: EMT models with nonlinear cores, residual flux, CT saturation and statistical breaker closing, including sympathetic inrush between parallel transformers.
• Controlled switching: feasibility studies, controller specification, residual-flux strategies and commissioning verification.
• Voltage sag and interconnection studies: energization impacts on transmission buses for solar, wind, battery and data center projects, against transmission owner requirements.
• CT specification and transient performance: IEEE C37.110 and C57.13 checks for external faults and long inrush tails.
• Misoperation analysis: event and COMTRADE review, root cause, corrective action plans and extent-of-condition reviews for NERC PRC-004.
• Commissioning support: pre-energization settings reviews, triggered oscillography, multiple-energization test plans and settings-basis updates.

Why Clients Choose Keentel

• Protection and system studies under one roof. Our protection engineers and EMT modelers work on the same problem, so settings reflect how the transformer and grid actually behave.
• US standards fluency. IEEE C37.91, C37.110, C57.12.00, C57.13 and NERC PRC-004 and PRC-005 applied correctly.
• Evidence-based settings. We use recorded energizations and statistical studies, not generic defaults.
• From 4 kV to 765 kV. Distribution substations, transmission autotransformers, generator step-up and main power transformers for inverter-based plants and large loads.
• Responsive senior engineers who stay with the project from study through energization.

Talk to Keentel Engineering

Planning to energize a new or replacement transformer, or investigating a differential trip on energization? Keentel Engineering can help you set the protection, plan the energization and prove the result.

• Phone: (813) 389-7871
• Email: contact@keentelengineering.com
• Book a 15-minute call: calendly.com/keentel-engineering/15min
• Web: keentelengineering.com

References

• IEEE Std C37.91, IEEE Guide for Protecting Power Transformers.
• IEEE Std C37.110, IEEE Guide for the Application of Current Transformers Used for Protective Relaying Purposes.
• IEEE Std C37.90, IEEE Standard for Relays, Relay Systems, and Control Devices Used for Protection and Control of Electric Power Apparatus.
• IEEE Std C37.2, IEEE Standard for Electrical Power System Device Function Numbers, Acronyms, and Contact Designations.
• IEEE Std C57.12.00, IEEE Standard for General Requirements for Liquid-Immersed Distribution, Power, and Regulating Transformers.
• IEEE Std C57.13, IEEE Standard Requirements for Instrument Transformers.
• IEEE Std C57.109, IEEE Guide for Liquid-Immersed Transformers Through-Fault-Current Duration.
• IEEE Std C57.142, IEEE Guide to Describe the Occurrence and Mitigation of Switching Transients Induced by Transformers, Switching Device, and System Interaction.
• IEEE Std 1453, IEEE Recommended Practice for the Analysis of Fluctuations on Electric Power Systems.
• IEEE Power System Relaying and Control Committee, technical reports on transformer protection and inrush.
• NERC Reliability Standard PRC-004, Protection System Misoperation Identification and Correction.
• NERC Reliability Standard PRC-005-6, Protection System, Automatic Reclosing, and Sudden Pressure Relaying Maintenance.
• Blackburn, J. L., and Domin, T. J., Protective Relaying: Principles and Applications, CRC Press.
• Horowitz, S. H., and Phadke, A. G., Power System Relaying, Wiley.

Disclaimer

This article is provided by Keentel Engineering for general technical education. It is not engineering advice for any specific project. The transformer data, saturation characteristics, residual fluxes, relay settings, simulation results and case studies are illustrative or composite.

Inrush behavior depends on the specific transformer, source, breaker and relay; protection settings must be based on project-specific studies, manufacturer data, utility protection philosophy and the current editions of the applicable standards, and sealed by a licensed professional engineer where required. Keentel Engineering accepts no liability for decisions made on the basis of this article without a project-specific engineering review.

© 2026 Keentel Engineering. All rights reserved. Written by Keentel Engineering.