9. Transformer and Reactor Protection

Introduction and Context

Power transformers and reactors are among the most valuable and strategically important assets in a power system. A transformer connects networks operating at different voltage levels, while a reactor controls current, voltage, reactive power, or transient overvoltages. Failure of either apparatus can interrupt supply, damage adjacent equipment, and create severe fire and safety hazards. Protection must therefore detect internal faults rapidly while remaining stable for external faults, energisation, through-faults, and permissible operating conditions.

Transformer protection is unusually demanding because the protected equipment is not electrically uniform. The primary and secondary currents differ in magnitude and phase, windings may be connected in different vector groups, and magnetising current is strongly nonlinear. Current transformers (CTs) can saturate during high-current external faults, producing differential spill current even when the transformer is healthy. In addition, conditions such as overfluxing, overheating, dissolved gas formation, and pressure rise may develop without producing large fault currents.

The protection system must therefore combine electrical, thermal, mechanical, and condition-based functions. The central design problem is selectivity: an internal fault should cause rapid tripping, whereas an external fault or temporary abnormal condition should result in restraint, alarm, or controlled action rather than unnecessary disconnection.

Learning Objectives

After studying this chapter, the reader should be able to:

1. Explain the operating principles and protection zones of transformers and reactors.

2. Design and assess biased differential protection, including ratio compensation, vector-group compensation, CT selection, and stability during CT saturation.

3. Distinguish restricted-earth-fault protection from conventional differential protection and identify suitable applications.

4. Explain magnetising inrush, sympathetic inrush, overfluxing, thermal stress, gas accumulation, and pressure protection.

5. Select protection functions for transformers and reactors according to rating, construction, grounding, and system importance.

6. Analyse practical protection misoperations and propose secure and dependable solutions.

The Protected Zone and the Faults That Matter

A transformer protection zone is normally bounded by CTs installed on each electrical connection. The zone includes the transformer windings and, depending on the design, parts of the bushings and connecting leads. A fault inside this zone is an internal fault; a fault outside it is an external or through fault. Correct CT location is essential because differential protection compares currents entering and leaving the zone.

Internal transformer faults include phase-to-earth, phase-to-phase, interturn, winding-to-core, and open-circuit faults. Earth faults near the neutral point may involve relatively low voltage and low current, particularly when the neutral is impedance grounded. Interturn faults can initially involve only a few turns and may not produce a large terminal current. Mechanical protection and sensitive earth-fault schemes are consequently important complements to phase differential protection.

External faults impose high through currents and severe electromechanical forces on the windings. The protection must remain stable during these events. Overloads, low system frequency, overvoltage, cooling failure, and repeated energisation may not be faults in the narrow sense, but they can produce damaging thermal or dielectric stress.

Reactors require a related but distinct approach. Shunt reactors may experience winding faults, turn-to-turn faults, core faults, overtemperature, and excessive current. Series reactors are subject to through-fault forces and insulation stress, while neutral grounding reactors require protection coordinated with the system earth-fault scheme. Reactor differential protection is often simpler than transformer differential protection because there is usually no ratio or phase displacement between terminals, but CT saturation and inrush-like transients can still influence security.

Differential Protection: The Primary Electrical Defence

Differential protection is based on Kirchhoff’s current law. For an ideal healthy transformer, the vector sum of currents entering the protected zone is zero after correction for transformation ratio and phase displacement. An internal fault produces a non-zero operating or differential current.

For a two-terminal apparatus, a simplified operating quantity is

\[

I_{op}=|I_1'+I_2'|

\]

where the currents have been referred to a common magnitude and phase basis. The restraining quantity is commonly related to the through current, for example

\[

I_{rest}=\frac{|I_1'|+|I_2'|}{2}.

\]

A biased or percentage-restrained relay operates when the differential current exceeds a characteristic such as

\[

I_{op}>I_{pickup}+S I_{rest},

\]

where \(S\) is the bias slope. Modern relays usually employ a two-slope or multi-slope characteristic. A lower slope provides sensitivity for internal faults at modest current, while a higher slope provides security during heavy external faults and CT saturation.

The use of bias is fundamental. During an external fault, one CT may saturate before the other. Its secondary current then becomes distorted and smaller than expected, creating apparent differential current. A high restraining current signals that the system is carrying a large through current, so the relay raises its operating threshold. Security must nevertheless be balanced against sensitivity to high-impedance internal faults.

Ratio and Phase Compensation

Transformer currents cannot be compared directly because the current ratio is determined by the transformer rating and winding voltages. Historically, relay connections and interposing CTs performed much of this compensation. Numerical relays generally perform it mathematically, but the engineer must still enter correct rated currents, CT ratios, vector group, and grounding configuration.

The transformer’s vector group introduces a phase displacement, commonly 30 degrees for many standard connections. A current entering a delta winding is not represented in the same phase relationship as current entering a star winding. The differential algorithm must rotate and combine the measured phasors according to the actual vector group. An incorrect vector-group setting can create substantial spill current under normal load and may result in nuisance tripping or blocked sensitivity.

Zero-sequence current also requires attention. In a grounded star–delta transformer, zero-sequence current on the star side may circulate within the delta or be prevented from appearing at the other terminals. Protection design must reflect the physical path of zero-sequence current. Modern numerical compensation can reproduce the required phase and zero-sequence behaviour, but it cannot correct incorrect wiring or incorrect configuration data.

CT Selection and Saturation

CTs for transformer differential protection should have adequate accuracy, burden capability, knee-point voltage, and transient performance. Their ratios should be selected so that relay inputs are compatible across all transformer terminals. CT secondary circuits must be short, correctly wired, securely earthed at one point, and protected against accidental open circuiting.

CT saturation is particularly likely during an external fault with a large asymmetrical current component. The DC offset drives the CT core toward saturation, causing secondary current distortion. Protection engineers assess stability using CT excitation characteristics, connected burden, remanence, fault current, and system X/R ratio. Percentage bias, saturation detection, waveform analysis, and adaptive restraint are complementary measures rather than substitutes for sound CT design.

A differential relay should not be set merely by increasing its pickup. Excessive pickup reduces sensitivity to internal faults, especially faults close to the neutral or faults involving limited energy. The preferred strategy is an appropriate CT class, correct wiring, adequate bias, and a relay algorithm designed for transient conditions.

Restricted Earth-Fault Protection

Restricted-earth-fault (REF) protection is a sensitive unit protection scheme for earth faults within a defined portion of a transformer winding. It compares the residual current from the three phase CTs with the neutral CT current. For an external earth fault, the two quantities balance; for an internal earth fault, they produce spill current.

REF can be high impedance or low impedance. In a high-impedance scheme, the CTs are connected in parallel and the relay circuit is designed so that the relay voltage remains below its stability limit during external faults. A stabilising resistor is commonly used. The design must consider the maximum secondary voltage generated by CT saturation and the relay burden. High-impedance REF requires closely matched CTs and careful wiring.

Low-impedance REF is implemented numerically using measured phase and neutral currents, usually with bias and saturation security. It is more flexible with respect to CT characteristics and can accommodate additional terminals, but the settings and directional logic must be carefully engineered.

REF is more sensitive than overall transformer differential protection for faults near the grounded neutral because the phase currents may be small while the neutral current is clear. It does not, however, protect phase-to-phase faults unless they also involve earth, nor does it normally cover the entire transformer unless CT placement and winding arrangement permit it.

Magnetising Inrush and Sympathetic Inrush

When a transformer is energised, the core flux is determined by the residual flux and the point on the voltage wave at which the breaker closes. If the resulting flux exceeds the normal operating range, the core saturates during part of the cycle. Magnetising current can then be several times rated current and may contain substantial second-harmonic content. Since this current enters through one terminal and does not leave through another, it appears as differential current.

A traditional restraint method blocks or restrains tripping when the ratio of second-harmonic current to fundamental current exceeds a selected threshold. Modern relays may use waveform asymmetry, fifth-harmonic behaviour, flux estimation, adaptive algorithms, or current-voltage analysis. No single criterion is universally reliable: harmonic levels vary with transformer design, CT performance, system impedance, and the severity of saturation.

Over-reliance on second-harmonic restraint can reduce sensitivity to some internal faults because fault current may also contain harmonics. Therefore, harmonic restraint should be coordinated with differential magnitude, negative-sequence quantities, waveform criteria, and secure fault detection.

Sympathetic inrush occurs when an already energised transformer experiences a temporary change in flux and current because another transformer is energised on the same network. The phenomenon arises from DC flux components and network coupling. It may last longer than ordinary energisation inrush and can cause apparent differential current or upstream overcurrent operation. Protection settings, energisation procedures, controlled switching, point-on-wave closing, and system studies should be considered where repeated misoperation occurs.

Overfluxing Protection

Overfluxing is associated with excessive volts-per-hertz, approximately represented by

\[

\frac{V}{f}.

\]

For a given number of turns, excessive voltage or reduced frequency increases core flux. The core may saturate, causing heating of the core, structural parts, and nearby metallic components. Overfluxing is principally a through-condition and may not produce a differential current suitable for fault protection.

A volts-per-hertz relay measures voltage and frequency, or estimates the ratio from the system quantities. Protection is often arranged with an alarm stage, a time-delayed trip stage, and sometimes an inverse-time characteristic. The thermal withstand of the transformer and the manufacturer’s V/Hz curve should determine settings. Generator step-up transformers are especially exposed during generator overspeed, load rejection, islanding, or incorrect excitation control.

Overfluxing protection must distinguish abnormal system conditions from instrument-transformer errors and transient voltage events. It should be coordinated with generator excitation limiters, automatic voltage regulators, underfrequency load shedding, and system operating procedures.

Thermal and Mechanical Protection

Transformer thermal damage is governed by winding hot-spot temperature, top-oil temperature, ambient temperature, loading history, and cooling performance. A thermal model estimates hot-spot temperature from load current and oil temperature. Protection may include temperature alarms, cooling-stage initiation, load reduction, and trip functions. Direct fibre-optic winding-temperature measurement can improve accuracy in large transformers.

Loss of cooling fans or pumps may not immediately justify tripping, but it reduces permissible loading. A staged response is generally preferable: alarm, start redundant cooling, reduce loading, and trip only when the thermal limit is approached or a cooling failure persists.

Pressure and mechanical protection respond to the physical consequences of internal faults. Sudden-pressure relays detect rapid pressure rise in oil-filled transformers and can trip quickly. Pressure-relief devices prevent tank rupture by releasing excessive pressure; they may provide an electrical alarm or trip contact. Conservator transformers may use a Buchholz relay in the pipe between the main tank and conservator. Slow gas accumulation produces an alarm, while a sudden oil surge caused by a serious internal fault produces a trip. Buchholz protection is not applicable to sealed transformers without a conservator in the conventional form.

Gas protection is valuable because arcing, overheating, and cellulose decomposition generate gases even when fault current is limited. Dissolved-gas analysis (DGA) complements, rather than replaces, electrical protection. Trends in hydrogen, methane, ethylene, acetylene, carbon monoxide, and carbon dioxide can indicate electrical or thermal deterioration. Interpretation should use recognized diagnostic methods and transformer-specific baselines; a single gas concentration should not automatically be treated as proof of a particular fault.

Reactor Protection Design

The reactor protection arrangement depends on whether the unit is a shunt, series, filter, arc-suppression, or neutral grounding reactor. Differential protection is suitable for many shunt reactors, using CTs on the line and neutral sides. Earth-fault and phase overcurrent elements provide backup. For air-core reactors, external magnetic fields and mechanical forces require attention to CT placement and physical clearances.

Shunt reactors can experience turn-to-turn faults that produce modest terminal-current changes. Sensitive differential, negative-sequence, neutral displacement, and mechanical protection may therefore be combined. Series reactors are normally protected against internal faults by differential or high-speed overcurrent schemes, while their through-fault withstand and bypass arrangements must be coordinated with system protection.

Protection of a neutral grounding reactor must be integrated with the neutral earthing philosophy. The relay should distinguish a fault inside the reactor from the earth faults that the reactor is intended to limit. Incorrect coordination can either leave the reactor exposed or disconnect healthy feeders unnecessarily.

A Practical Protection Architecture

A typical important oil-immersed transformer may use the following coordinated functions:

• biased phase differential protection for internal phase and earth faults;
• REF for sensitive earth faults on grounded windings;
• backup phase and earth overcurrent protection;
• overfluxing protection based on volts per hertz;
• winding and oil temperature alarms and trips;
• Buchholz, sudden-pressure, and pressure-relief indications where applicable;
• oil-level, cooling-failure, and tap-changer protection;
• breaker-failure, lockout, and trip-circuit supervision;
• DGA and online monitoring for condition assessment.

Protection functions should not all be assigned the same action. A gas accumulation alarm, for example, normally initiates investigation rather than immediate tripping. A high-speed differential trip generally operates a lockout relay, opens all relevant breakers, blocks automatic re-energisation, and records the event for investigation.

Coordination includes operating time, sensitivity, CT performance, breaker capability, transformer damage curves, system fault levels, and maintenance practices. Commissioning must verify phase relationships, CT polarity, ratio, insulation, trip paths, relay settings, event recording, and interlocking. A correct relay setting cannot compensate for reversed CT polarity or an untested trip circuit.

Illustrative Case Study: Differential Operation During Energisation

Background

A 220/33 kV power transformer was commissioned with numerical biased differential protection. The transformer used a grounded star connection on the high-voltage side and a delta connection on the low-voltage side. The transformer had a standard phase displacement, and CTs were installed at both terminals.

Situation

During initial energisation, the differential relay issued a trip command after approximately one cycle. No visible damage was found, insulation tests were satisfactory, and the transformer had not been connected to the 33 kV load. The event recorder showed a large differential current on one phase with a strongly asymmetric waveform.

Analysis

The first hypothesis was magnetising inrush. However, the relay had been configured with the wrong vector group, so normal phase displacement was not correctly compensated. In addition, the second-harmonic restraint threshold had been set without reviewing the manufacturer’s expected inrush waveform. A polarity check also revealed that one low-voltage CT phase had been transposed during wiring.

The incorrect vector-group setting created differential current under balanced load conditions, while the CT phase transposition increased the apparent mismatch. During energisation, the asymmetric inrush exceeded the relay pickup before the restraint logic could provide security.

Corrective Approach

The commissioning team verified CT polarity and phase identity using primary injection and compared measured phasors with the transformer nameplate vector group. The numerical relay was configured with the correct ratio, rated currents, vector group, zero-sequence treatment, and inrush restraint. Stability was tested by secondary injection for external-fault conditions, and energisation was repeated under controlled conditions.

Outcome

The transformer energised without an unwanted trip. Subsequent tests confirmed that the relay operated for simulated internal faults and remained stable for external-fault currents within the available test capability.

Learning Points

The case demonstrates that differential protection is a complete measurement system, not merely a relay setting. Vector-group data, CT polarity, ratio compensation, inrush logic, and commissioning tests are equally important. It also shows why a trip during energisation should not automatically be attributed to a defective relay or a faulty transformer.

Challenges, Limitations, and Common Errors

The most frequent errors are incorrect CT polarity, wrong vector-group entry, inadequate CT class, unrecognised zero-sequence paths, and failure to include all transformer terminals in the protected zone. Another risk is using settings copied from a similar transformer without checking rating, impedance, grounding, tap range, fault level, and CT burden.

Protection may be less sensitive to high-resistance and interturn faults than to solid terminal faults. Mechanical protection can be slow for incipient deterioration and may require periodic testing. DGA interpretation is probabilistic and can be affected by oil handling, sampling technique, and historical contamination. Online sensors introduce communication, calibration, cybersecurity, and data-quality concerns.

Trip logic should be fail-safe but not indiscriminate. Unnecessary transformer trips impose major operational and financial consequences, while failure to trip can cause catastrophic damage. Independent trip paths, lockout relays, redundant DC supplies, supervision, and disciplined maintenance are therefore justified for critical installations.

Digital Monitoring and Emerging Practice

Modern transformer protection increasingly combines high-speed protection with asset-health monitoring. Digital relays provide disturbance records, synchrophasor-compatible measurements, harmonic information, and self-supervision. Online DGA, moisture sensors, bushing power-factor monitoring, fibre-optic temperature measurement, and tap-changer monitoring support predictive maintenance.

Analytics can identify deviations from a transformer’s normal thermal, gas, or loading pattern. Machine-learning tools may assist anomaly detection, but their outputs should support—not replace—engineering judgement, validated diagnostic models, and deterministic trip protection. Cybersecurity is now part of protection engineering: authentication, network segmentation, secure configuration management, time synchronisation, and controlled remote access are necessary for intelligent electronic devices.

Best-Practice Design Principles

Effective transformer and reactor protection follows several principles:

1. Define the protection zone and fault types before selecting relay functions.

2. Obtain verified nameplate, vector-group, grounding, impedance, CT, and cooling data.

3. Use differential bias and CT performance studies to achieve both security and sensitivity.

4. Treat REF, gas, pressure, thermal, and overfluxing functions as complementary protections.

5. Coordinate alarm, cooling, load reduction, trip, and lockout actions according to damage mechanisms.

6. Test the complete chain from primary quantities or realistic secondary injection to breaker operation.

7. Record and investigate every operation using event files, DGA, inspection, and system information.

8. Review settings after network changes, transformer replacement, CT changes, or modifications to grounding.

Chapter Summary

Transformer and reactor protection must detect internal electrical faults rapidly while remaining stable for external faults, energisation, overload, and system disturbances. Biased differential protection is the principal electrical scheme, relying on correct CT ratios, polarity, vector-group compensation, zero-sequence treatment, and saturation security. REF provides enhanced sensitivity for grounded-winding earth faults. Magnetising and sympathetic inrush require restraint or recognition so that energisation does not imitate an internal fault. Overfluxing protection responds to excessive volts per hertz, while thermal, gas, sudden-pressure, Buchholz, and pressure-relief functions address heating, insulation deterioration, and mechanical consequences. Reactor protection is adapted to the reactor type and grounding role. Reliable performance depends on sound design, accurate configuration, commissioning, maintenance, condition monitoring, and disciplined trip logic.

Key Terms

Differential protection: Unit protection that compares currents entering and leaving a defined zone.

Biased differential: Differential protection whose operating threshold increases with restraining current.

Operating current: The differential or spill current that tends to operate a relay.

Restraining current: A quantity representing through current and used to improve security.

Vector group: Transformer designation specifying winding connections and phase displacement.

Ratio compensation: Adjustment that refers terminal currents to a common magnitude basis.

CT saturation: Nonlinear CT operation in which the secondary current no longer accurately reproduces primary current.

Magnetising inrush: Transient current drawn during transformer energisation because of core saturation.

Sympathetic inrush: Inrush-like current induced in an energised transformer when another transformer is energised nearby.

Restricted earth fault: Sensitive unit protection for earth faults within a specified winding zone.

Overfluxing: Excessive magnetic flux caused mainly by high volts per hertz.

Buchholz relay: Gas- and oil-surge-operated protection used on suitable oil-filled transformers with conservators.

Sudden-pressure relay: Device detecting rapid internal pressure rise in a transformer tank.

Dissolved-gas analysis: Diagnostic examination of gases dissolved in transformer oil.

Hot-spot temperature: The highest local winding temperature used to assess insulation ageing.

Shunt reactor: Reactor connected in parallel to absorb reactive power and control voltage.

Series reactor: Reactor connected in series to limit current or modify system behaviour.

Lockout relay: Latching relay that trips associated breakers and prevents automatic reclosing until reset.

Protected zone: Apparatus region bounded by the CTs associated with a unit protection scheme.

Volts-per-hertz relay: Protection that detects flux-related stress from excessive voltage-to-frequency ratio.

Review Questions

Short Answer Questions

1. Why is percentage bias used in transformer differential protection?

2. What information is represented by a transformer vector group?

3. Why can magnetising inrush cause differential relay operation?

4. What additional sensitivity does REF provide?

5. State two physical conditions detected by mechanical transformer protection.

Long Answer Questions

1. Explain the design and operation of biased differential protection, including CT saturation and restraint.

2. Compare high-impedance and low-impedance REF schemes.

3. Discuss the causes, characteristics, and protection treatment of magnetising and sympathetic inrush.

4. Explain overfluxing protection and its coordination with voltage and frequency control.

5. Develop a coordinated protection scheme for a large oil-immersed power transformer.

Analytical / Application Questions

1. A transformer differential relay trips during energisation but remains stable during load. Identify the investigations and tests required before changing the relay pickup.

2. An external earth fault causes differential spill current only at high fault current. Analyse the likely role of CT saturation and propose design remedies.

3. A transformer shows rising dissolved acetylene, repeated Buchholz alarms, and normal load current. Decide how protection, condition assessment, and operational actions should be coordinated.

22. Multiple Choice Questions with Answers

Question 1

What is the principal reason for using percentage bias in transformer differential protection?

A. To increase sensitivity equally for all load currents

B. To maintain security during high through-fault currents and CT saturation

C. To eliminate the need for CT ratio matching

D. To detect overfluxing directly

Correct Answer: B. To maintain security during high through-fault currents and CT saturation

Question 2

A wrong transformer vector-group setting in a numerical differential relay is most likely to produce what effect?

A. Failure of the transformer cooling system

B. Incorrect phase compensation and spill current during healthy operation

C. Removal of all zero-sequence current from the system

D. Reduced transformer oil pressure

Correct Answer: B. Incorrect phase compensation and spill current during healthy operation

Question 3

Which condition most directly causes magnetising inrush?

A. A high-resistance external earth fault

B. Core flux exceeding the normal range after energisation

C. Loss of communication with the relay

D. Low oil level in the conservator

Correct Answer: B. Core flux exceeding the normal range after energisation

Question 4

REF protection is particularly valuable because it can detect:

A. All external phase faults on adjacent feeders

B. Sensitive earth faults within a defined grounded-winding zone

C. Only overfrequency conditions

D. Mechanical vibration in air-core reactors

Correct Answer: B. Sensitive earth faults within a defined grounded-winding zone

Question 5

During a heavy external fault, one CT saturates and produces a distorted secondary current. The likely protection consequence is:

A. Apparent differential current requiring bias and CT-saturation security

B. Complete elimination of differential current

C. Automatic correction of the transformer vector group

D. Operation of the Buchholz relay only

Correct Answer: A. Apparent differential current requiring bias and CT-saturation security

Question 6

Overfluxing protection primarily responds to excessive:

A. Current-to-impedance ratio

B. Voltage-to-frequency ratio

C. Oil-to-gas ratio

D. Neutral-to-phase resistance

Correct Answer: B. Voltage-to-frequency ratio

Question 7

Which device conventionally detects gas accumulation and sudden oil movement in a conservator-type transformer?

A. Buchholz relay

B. Distance relay

C. Power-factor controller

D. Underfrequency relay

Correct Answer: A. Buchholz relay

Question 8

A transformer is healthy, but energisation of a nearby transformer causes a temporary inrush-like current. This is called:

A. Restricted earth fault

B. Sympathetic inrush

C. Overfluxing trip

D. CT remanence compensation

Correct Answer: B. Sympathetic inrush

Question 9

Which protection arrangement is most appropriate for a critical oil-immersed transformer?

A. Differential protection alone

B. Only phase overcurrent protection

C. Coordinated electrical, thermal, gas, pressure, and overfluxing functions

D. Only a manual inspection programme

Correct Answer: C. Coordinated electrical, thermal, gas, pressure, and overfluxing functions

Question 10

For a neutral grounding reactor, protection must primarily be coordinated with:

A. The system earthing and earth-fault philosophy

B. The transformer vector group on an unrelated feeder

C. The generator turbine lubrication system

D. The lighting load schedule

Correct Answer: A. The system earthing and earth-fault philosophy