Transformer Differential Protection (87T): Slope & Harmonics
Match transformer CT ratios and vector-group phase shift, plot a dual-slope 87T characteristic, and check through faults, internal faults and harmonic blocking.
Transformer differential protection compares currents on both sides of a power transformer after putting them on a common base. This 87T calculator helps protection engineers and students explore CT ratio matching, vector-group compensation, percentage differential pickup and harmonic blocking. It is useful for reviewing the logic behind a setting study or understanding a secondary-injection test.
Enter the transformer three-phase MVA rating, both rated line-to-line voltages and the installed CT primary and secondary ratings. The tool calculates rated line currents and the CT secondary currents at rated transformer load. These secondary values are the matching taps: dividing each measured current by its own tap gives comparable per-unit currents even when the two CT ratios differ.
The phase model is deliberately restricted to balanced ABC positive sequence with star-connected CT secondaries. Both measured phasors point into the protected zone. A healthy through current therefore cancels after compensation. The LV voltage displacement follows the nameplate clock convention, and the tool rotates the LV current by the opposite displacement. Clock 11 represents a 30° LV lead. This balanced rotation does not implement the three-phase compensation matrices or zero-sequence removal required for earth faults.
Choose one of four illustrative operating cases or enter measured fundamental CT secondary RMS magnitudes, phase angles and differential harmonic percentages. The default through-fault example uses 6 pu on the HV side and a 10% LV under-reading. The internal-fault example has 4 pu and 2 pu feeding inward. The comparison table evaluates both examples using the same settings, so a change in slope or compensation has an immediately visible consequence.
The operating characteristic uses differential current equal to the magnitude of the phasor sum and restraint current equal to half the sum of the magnitudes. A minimum pickup floor and two continuous rising slope segments define the threshold. The chart places the selected operating point against this curve. A positive current margin indicates pickup by the percentage characteristic; the final decision also considers harmonic blocking.
Second-harmonic inrush restraint is represented here by a low-set blocking threshold. Fifth-harmonic overexcitation blocking is separate. The harmonic chart compares the supplied ratios with their settings; it does not infer harmonics from a waveform or diagnose the cause of distortion. An optional unrestrained high-set element bypasses both blocks. Its default is disabled.
Settings shown are examples. Check against IEEE C37.91-2021, the standard in force and the relay manual in force before applying a method to an installation. Actual relays differ in restraint definition, characteristic geometry, harmonic sharing, filtering, timers and saturation detection. This calculator does not model CT transient saturation, unbalanced faults, restricted earth fault protection, tap-changer tracking, breaker operation or protection coordination. Fifth-harmonic blocking is not ANSI 24 V/Hz protection. Protection settings and commissioning belong to a qualified engineer.
How it is calculated
Match the CT currents
For each winding, rated line current is . The CT ratio is and the rated-load secondary tap is . A measured secondary RMS phasor divided by its own tap is on the common transformer per-unit base. ABB RET 54_/Diff6T Application and Setting Guide, 1MRS 755916 (2005), attached PDF page 8, gives the rated-current and ratio-matching method and the 25 MVA example.
Compensate balanced phase displacement
All current arrows point into the zone. For balanced ABC positive sequence, the LV nameplate voltage lag is for clock number . Rotate the LV inward-current phasor by . Ideal through currents then oppose and cancel. This is a balanced positive-sequence equivalent of vector-group matching, not a full three-phase relay compensation matrix. The preset through current has an LV angle ; the preset internal contribution has angle . Star-connected CT secondaries are assumed.
Calculate differential and average restraint
With compensated matched phasors , and . Ehsan et al., Energies 14 (2021) 8447, DOI 10.3390/en14248447, attached PDF page 9, describes this average-restraint convention. Other relays use different restraint definitions.
Apply this explicit continuous characteristic
Slopes entered in percent are divided by 100. The two rising segments meet continuously at ; the minimum pickup is a floor. The input condition keeps the knee above that floor. Percentage pickup requires strictly . This selectable example curve is not a reproduction of any manufacturer's entire relay characteristic.
Apply harmonic blocking and optional high set
Supplied and are ratios of compensated differential RMS components. An enabled harmonic block asserts at or above its threshold while . Either block inhibits the low-set percentage element. This implements inrush restraint as blocking, not additive harmonic restraint. If enabled, the unrestrained element trips for regardless of either block or slope. The result is a static logical decision, not a trip time. ABB attached PDF pages 11–13 discusses the distinction and limitations; Ehsan et al. page 16 reports harmonic injection tests.
Assumptions
- Two-winding, balanced ABC positive-sequence equivalent with ideal ratio scaling and star-connected CT secondaries; no delta-connected CT circuits.
- Nameplate MVA and rated voltages define both winding bases. No measured tap-position compensation, automatic winding identification or polarity detection.
- The four operating presets are author-defined teaching examples: through fault 6 / 5.4 pu; internal fault 4 + 2 pu; inrush 3 / 0 pu with 20% second harmonic; overexcitation 1.2 / 0 pu with 40% fifth harmonic. Other preset harmonic ratios are 2% second and 1% fifth.
- The comparison table uses the through and internal examples with the selected compensation and protection settings, and 2% / 1% harmonics.
- Manual inputs represent prefiltered fundamental phasors and already computed differential harmonic ratios. Ratios are not calculated from winding harmonic percentages or from waveforms.
- Equality at the percentage or high-set threshold does not trip; equality at a harmonic threshold blocks. No hysteresis, harmonic sharing or time delay is modeled.
Where it applies
Use for balanced current matching, characteristic studies and interpretation of balanced three-phase injection tests. Clock rotation alone cannot represent unbalanced or earth-fault compensation: zero-sequence removal, negative-sequence transformation and the actual relay matrices are outside this model. It is not suitable for three-winding transformers, differential zones with multiple breakers or autotransformer neutral/tertiary studies.
CT under-reading in the through-fault preset is a static magnitude error, not a CT saturation simulation or a guarantee of security. CT excitation, burden, remanence, fault X/R, waveform filtering and adaptive restraint can change actual performance. No CT class or saturation adequacy is checked.
Fifth-harmonic blocking neither measures V/Hz nor protects the transformer from overexcitation damage; separate ANSI 24 protection and manufacturer withstand data are needed. Harmonics can also occur during faults, and inrush can have low second-harmonic content. The high-set bypass and all harmonic settings require the relay's own application rules.
IEEE C37.91-2021 was consulted for terminology and scope. Check against the standard in force and the relay manual in force. No standard tables or figures are reproduced; this calculator does not certify compliance.
References
- IEEE Std C37.91-2021, IEEE Guide for Protecting Power Transformers. DOI: https://doi.org/10.1109/IEEESTD.2021.9471045. Consult the licensed edition and the standard in force.
- ABB, RET 54_/Diff6T Differential Protection Application and Setting Guide, 1MRS 755916 (2005), sections 3.1.1–3.1.8. https://library.e.abb.com/public/76c0ba85c2011f4fc225707b0050fbb4/application_settingguide_RET54_Diff6T_ENa.pdf
- U. Ehsan et al., A Detailed Testing Procedure of Numerical Differential Protection Relay for EHV Auto Transformer, Energies 14 (2021) 8447. https://doi.org/10.3390/en14248447. Open copy: https://psecommunity.org/wp-content/plugins/wpor/includes/file/2303/LAPSE-2023.16978-1v1.pdf
Inputs
| Input | Symbol | Unit | Default |
|---|---|---|---|
| Transformer rating Common three-phase apparent-power base for both windings. | S_n | MVA | 25 |
| HV rated line-to-line voltage | U_H | kV | 110 |
| LV rated line-to-line voltage | U_L | kV | 21 |
| HV CT rated primary current | CT_Hp | A | 300 |
| HV CT rated secondary current Enter 1 A or 5 A as installed. CT secondaries are assumed star connected. | CT_Hs | A | 1 |
| LV CT rated primary current | CT_Lp | A | 1000 |
| LV CT rated secondary current | CT_Ls | A | 1 |
| Nameplate vector-group clock number Balanced ABC positive sequence only. LV voltage lags HV by clock × 30°; clock 11 is a 30° lead. | clock_n | 11 | |
| Apply LV phase compensation Rotates LV inward current by +clock × 30°. Turn off to study a compensation error. | comp_on | true | |
| Operating case Through-fault preset: 6 pu HV, 5.4 pu LV (10% LV under-reading). Presets are illustrative balanced cases, not field measurements. The comparison checks the through and internal cases. | case_mode | Through fault: 6 / 5.4 pu | |
| HV fundamental CT secondary RMS current | I_Hm | A | 2.624319 |
| HV inward-current phase angle | a_Hm | deg | 0 |
| LV fundamental CT secondary RMS current | I_Lm | A | 3.711538 |
| LV inward-current phase angle before compensation Both currents are positive into the zone. Ideal through current has LV angle = HV angle + 180° − clock × 30°. | a_Lm | deg | 210 |
| Second harmonic / differential fundamental (%) Ratio of compensated differential RMS components, not the harmonic percentage of a winding current. No waveform or DFT calculation is performed. | h2_m | 2 | |
| Fifth harmonic / differential fundamental (%) | h5_m | 1 | |
| Minimum differential pickup (pu) Example setting on the transformer current base. | I_pick | 0.3 | |
| Slope 1 (%) | slope_1 | 30 | |
| Slope 2 (%) Example slopes use average restraint. Do not transfer settings from a relay using sum or maximum restraint. | slope_2 | 80 | |
| Slope transition restraint current (pu) | R_knee | 3 | |
| Enable second-harmonic inrush block | block_2_on | true | |
| Second-harmonic blocking threshold (%) | H2_set | 15 | |
| Enable fifth-harmonic overexcitation block Study option only. Confirm suitability from transformer data and the relay manual. This does not protect against overexcitation. | block_5_on | true | |
| Fifth-harmonic blocking threshold (%) | H5_set | 35 | |
| Enable unrestrained high-set element If enabled, fundamental differential above this threshold trips regardless of slope and harmonic blocks. Relay implementations vary. | high_on | false | |
| Unrestrained high-set pickup (pu) | I_high | 8 |
Results
With the default inputs above, the calc gives:
| Result | Value |
|---|---|
| Selected-case decision | RESTRAINED · below characteristic |
| Differential current (pu) | 0.6 |
| Average restraint current (pu) | 5.7 |
| Differential pickup threshold (pu) | 3.06 |
| Margin above characteristic (pu) | -2.46 |
| Second-harmonic block active | no |
| Fifth-harmonic block active | no |
| HV matching tap: secondary at rated load | 0.4374 A |
| LV matching tap: secondary at rated load | 0.6873 A |
| Applied LV phase rotation | 330 deg |
Checked against sources
Each case below was calculated by this calc and agrees with the value from its source (within the tolerance where one is shown).
ABB 25 MVA example: HV rated current and ratio matching
Source: ABB RET 54_/Diff6T, 1MRS 755916 (2005), attached PDF page 8 (printed page 8), 25 MVA / 110 kV worked example. Published rounding uses 1.732; the calculator uses sqrt(3).
Inputs: the default inputs above.
| Result | Source | This calc |
|---|---|---|
| Rated HV line current | 131.2 A | 131.2 A |
| HV CT matching factor relative to its rated secondary | 2.287 | 2.286 |
ABB 25 MVA example: LV rated current and ratio matching
Source: ABB RET 54_/Diff6T, 1MRS 755916 (2005), attached PDF page 8, 25 MVA / 21 kV worked example.
Inputs: the default inputs above.
| Result | Source | This calc |
|---|---|---|
| Rated LV line current | 687.3 A | 687.3 A |
| LV CT matching factor relative to its rated secondary | 1.455 | 1.455 |
Published balanced HV-only injection: 1 pu differential, 0.5 pu restraint
Source: Ehsan et al., Energies 14 (2021) 8447, DOI 10.3390/en14248447, PDF page 9 / Fig. 6a: balanced three-phase 0.35 A HV injection, LV zero, at 160 MVA, 220/132 kV, CT 1200/1. This verifies a balanced equivalent, not an autotransformer's neutral or tertiary model. Published rounded 420 A base gives exactly 1 and 0.5; retaining sqrt(3) gives a small rounding residual.
Inputs: Operating case = Enter measured CT phasors and harmonics; Transformer rating = 160 MVA; HV rated line-to-line voltage = 220 kV; LV rated line-to-line voltage = 132 kV; HV CT rated primary current = 1,200 A; LV CT rated primary current = 1,200 A; Nameplate vector-group clock number = 0; LV inward-current phase angle before compensation = 180 deg; Second harmonic / differential fundamental (%) = 0; Fifth harmonic / differential fundamental (%) = 0; Minimum differential pickup (pu) = 0.15; Slope 2 (%) = 60; Slope transition restraint current (pu) = 2; HV fundamental CT secondary RMS current = 0.35 A; LV fundamental CT secondary RMS current = 0 A; the others as above.
| Result | Source | This calc |
|---|---|---|
| Differential current (pu) | 1 ± 0.002 | 1 |
| Average restraint current (pu) | 0.5 ± 0.002 | 0.5001 |
| Static trip flag: 1 = trip, 0 = no trip | 1 | 1 |
Published five-times-load balanced through-fault injection remains stable
Source: Ehsan et al., DOI 10.3390/en14248447, Table 8, PDF page 15: 1.75 A HV, 2.915 A LV, opposite compensated phases; published calculated D=0 and R=5 pu, observed D=0.012 and R=5.04 pu. The paper rounds rated currents and taps; unrounded nameplate bases leave about 0.00286 pu spill, retained here within 0.004 pu absolute tolerance. No CT transient model is inferred from the measurement.
Inputs: Operating case = Enter measured CT phasors and harmonics; Transformer rating = 160 MVA; HV rated line-to-line voltage = 220 kV; LV rated line-to-line voltage = 132 kV; HV CT rated primary current = 1,200 A; LV CT rated primary current = 1,200 A; Nameplate vector-group clock number = 0; LV inward-current phase angle before compensation = 180 deg; Second harmonic / differential fundamental (%) = 0; Fifth harmonic / differential fundamental (%) = 0; Minimum differential pickup (pu) = 0.15; Slope 2 (%) = 60; Slope transition restraint current (pu) = 2; HV fundamental CT secondary RMS current = 1.75 A; LV fundamental CT secondary RMS current = 2.915 A; the others as above.
| Result | Source | This calc |
|---|---|---|
| Differential current (pu) | 0 ± 0.004 | 0.002858 |
| Average restraint current (pu) | 5 | 5 |
| Static trip flag: 1 = trip, 0 = no trip | 0 | 0 |
Published percentage-pickup injection point B trips
Source: Ehsan et al., DOI 10.3390/en14248447, Table 9 point B, PDF page 16. Measured D=0.18 pu and R=0.404 pu; the input secondary values are rounded and the experimental readings include measurement error. Characteristic set to the paper's 0.15 pu / 30% / 60% / 2 pu values. This point demonstrates low-set operation, not high-set or saturation algorithms.
Inputs: Operating case = Enter measured CT phasors and harmonics; Transformer rating = 160 MVA; HV rated line-to-line voltage = 220 kV; LV rated line-to-line voltage = 132 kV; HV CT rated primary current = 1,200 A; LV CT rated primary current = 1,200 A; Nameplate vector-group clock number = 0; LV inward-current phase angle before compensation = 180 deg; Second harmonic / differential fundamental (%) = 0; Fifth harmonic / differential fundamental (%) = 0; Minimum differential pickup (pu) = 0.15; Slope 2 (%) = 60; Slope transition restraint current (pu) = 2; HV fundamental CT secondary RMS current = 0.171 A; LV fundamental CT secondary RMS current = 0.18 A; the others as above.
| Result | Source | This calc |
|---|---|---|
| Differential current (pu) | 0.18 ± 0.005 | 0.18 |
| Average restraint current (pu) | 0.404 ± 0.01 | 0.3987 |
| Static trip flag: 1 = trip, 0 = no trip | 1 | 1 |
Published second-slope injection point F trips
Source: Ehsan et al., DOI 10.3390/en14248447, Table 9 point F, PDF page 16. Secondary injections and measured readings verify second-slope operation. Fig. 8 labels the ideal differential coordinate 1.25 pu, while Table 9 reports measured 1.26 pu; this check follows the measured result with allowance for rounding and measurement error, retaining the stated equations. The measured bias is about 1% above the value reconstructed from the rounded injection currents; a 0.04 pu absolute tolerance covers that measurement difference.
Inputs: Operating case = Enter measured CT phasors and harmonics; Transformer rating = 160 MVA; HV rated line-to-line voltage = 220 kV; LV rated line-to-line voltage = 132 kV; HV CT rated primary current = 1,200 A; LV CT rated primary current = 1,200 A; Nameplate vector-group clock number = 0; LV inward-current phase angle before compensation = 180 deg; Second harmonic / differential fundamental (%) = 0; Fifth harmonic / differential fundamental (%) = 0; Minimum differential pickup (pu) = 0.15; Slope 2 (%) = 60; Slope transition restraint current (pu) = 2; HV fundamental CT secondary RMS current = 1.269 A; LV fundamental CT secondary RMS current = 1.384 A; the others as above.
| Result | Source | This calc |
|---|---|---|
| Differential current (pu) | 1.26 ± 0.015 | 1.253 |
| Average restraint current (pu) | 3.031 ± 0.04 | 3 |
| Static trip flag: 1 = trip, 0 = no trip | 1 | 1 |
Related Resources
Related courses
IEC Electrical Schematics and Industrial Control Panels
Basic Electrical Controls & Diagram Reading

Responses