Power Systems
Ferranti Effect: Why Voltage Rises at the Far End of a Lightly-Loaded Line
Energize a 400 km, 400 kV overhead line with the far end open, and the receiving-end voltage can climb roughly 5 percent above the sending end; do the same on a long EHV cable and the rise can exceed 10 percent — enough to threaten insulation and trip protection before a single megawatt has been delivered. This counterintuitive voltage gain along a wire that has resistance and should only ever lose voltage is the Ferranti effect.
The Ferranti effect is the rise of receiving-end voltage above sending-end voltage on a long AC transmission line operated at no load or light load. It arises because the line's distributed shunt capacitance draws a leading (capacitive) charging current, and that current flowing through the line's series inductance produces a voltage rise that adds to, rather than subtracts from, the source voltage.
- TypePower-system voltage-rise phenomenon
- Named afterSebastian Ziani de Ferranti
- First observed1887, London 10 kV AC cable system
- Key equationVr/Vs = 1/cos(βl); ΔV ≈ ½·ω²·L·C·l²·Vr
- Typical magnitude~5% per 300 km (OHL); >10% on long EHV cables
- Standard mitigationShunt reactors at line/receiving end
Interactive visualization
Press play, or step through manually. The visualization is yours to drive — try it before reading on.
Watch the 60-second explainer
A condensed visual walkthrough — narrated, captioned, under a minute.
What It Is and Where It Shows Up
The Ferranti effect is the phenomenon whereby the voltage at the receiving (far) end of a long AC transmission line becomes higher than the voltage at the sending (source) end when the line is open-circuited or lightly loaded. It was first documented by British engineer Sebastian Ziani de Ferranti around 1887 during the commissioning of his pioneering 10,000-volt AC underground cable system supplying central London from Deptford power station.
It matters most for:
- Long EHV overhead lines (220 kV, 400 kV, 765 kV) more than ~200–300 km, especially at night when load drops.
- HVAC underground and submarine cables, where shunt capacitance per km is 20–40× that of overhead lines, so even short cables show a marked rise.
- Line energization and load-rejection events, when a line is suddenly charged or loses its load and the far end floats up.
Because insulation, surge arresters, and generator excitation systems are all rated around nominal voltage, an uncontrolled Ferranti rise is a genuine operational hazard, not just a textbook curiosity.
How It Works: The Mechanism and Derivation
Model a line as a distributed series impedance (resistance R and inductance L per km) and distributed shunt capacitance C per km. At no load the receiving end draws no useful current, but the shunt capacitance still draws a leading charging current I_c ≈ jωC·V. This current is capacitive, so it leads the voltage by 90°.
When that leading current flows back through the series inductance, the inductive voltage drop jωL·I_c is oriented so it adds to the source phasor rather than subtracting. Physically, a capacitive current through an inductance produces a voltage rise. For a lossless line the exact long-line solution gives:
- Vs = Vr·cos(βl), hence Vr/Vs = 1/cos(βl),
where β = ω·√(L·C) is the phase constant (rad/km) and l is line length. Since cos(βl) < 1, Vr > Vs. Expanding for small βl gives the handy approximation ΔV = Vr − Vs ≈ ½·ω²·L·C·l²·Vr — the rise grows with the square of line length and with frequency squared.
Key Quantities and a Worked Example
Define the symbols: ω = 2πf (rad/s), L series inductance/km (H/km), C shunt capacitance/km (F/km), l length (km), β = ω√(LC) phase constant (rad/km), v = 1/√(LC) propagation velocity (≈ 3×10⁵ km/s for OHL).
Worked example — 300 km, 50 Hz overhead line. Typical values: L ≈ 1.0 mH/km, C ≈ 0.011 µF/km.
- ω = 2π·50 = 314.2 rad/s.
- β = ω√(LC) = 314.2·√(1.0e-3 · 0.011e-6) = 314.2·√(1.1e-11) ≈ 1.04×10⁻³ rad/km.
- βl = 1.04×10⁻³ · 300 ≈ 0.313 rad (17.9°).
- Vr/Vs = 1/cos(0.313) = 1/0.951 ≈ 1.051 → a 5.1% rise.
The approximation ½·(βl)² = ½·(0.313)² = 0.049 → 4.9% agrees well. For a 765 kV, 500 km line the same math yields roughly a 15–20% rise, which is why such lines are essentially never operated open without reactive compensation.
Managing It in Practice
Because the effect is driven by uncompensated line charging (reactive) power, mitigation means absorbing that capacitive VARs. Standard measures:
- Shunt reactors — the workhorse fix. An inductive reactor connected line-to-neutral at the receiving end (and often mid-line) draws lagging current that cancels the capacitive charging current. Reactors are commonly sized to compensate 60–80% of the line's charging MVAR (Qc = ωC·l·V²).
- Switched/variable compensation — SVCs and STATCOMs absorb VARs dynamically as load varies through the day, replacing fixed reactors on modern lines.
- Operational rules — avoid energizing long lines from the weak (open) end; keep some minimum load; use synchronous condensers or under-excited generators to absorb reactive power.
Grid codes (e.g., IEC/IEEE steady-state limits and national codes) typically cap steady voltage at +5% to +10% of nominal, so a designer computes the no-load rise and adds enough reactive absorption to stay inside that window across all load conditions, including light-load nights.
How It Relates to Other Line Phenomena
The Ferranti effect is one member of a family of long-line behaviors, and it's easy to confuse with its cousins:
- Voltage regulation / IR drop — the opposite, everyday case: under normal load, series R and X cause the receiving voltage to sag. Ferranti is the light-load exception where charging current flips the sign.
- Surge impedance loading (SIL) — at load equal to SIL = V²/Zc (Zc = √(L/C)), the line's inductive and capacitive VARs balance and the voltage profile is flat. Below SIL you get Ferranti rise; above it, voltage sag.
- Line resonance / quarter-wave effect — as βl → π/2 (~1500 km at 50 Hz), cos(βl) → 0 and Vr/Vs → ∞. Ferranti is the low-βl portion of that same resonance curve.
- Corona and skin effect — loss mechanisms, unrelated to the capacitive rise but relevant to the same conductors.
Understanding SIL is the key intuition: Ferranti is simply what happens when a line carries far less than its natural load.
Failure Modes, Trade-offs, and Significance
Left unmanaged, the Ferranti rise causes real damage:
- Insulation overstress — sustained overvoltage ages cable and transformer insulation and can flash over bushings; on cables the effect is worst because of huge shunt capacitance.
- Transformer saturation — overvoltage pushes core flux past the knee, drawing large magnetizing currents, generating harmonics and overheating.
- Protection and equipment trips — surge arresters and overvoltage relays operate, and generator AVRs fight the rise.
The trade-offs of the standard fix are also real: fixed shunt reactors that tame the no-load rise become a liability under heavy load, where they worsen voltage sag and eat into transfer capacity — hence the move to switchable or dynamic (SVC/STATCOM) compensation.
Its significance is enduring: the Ferranti effect sets a practical ceiling on how long an uncompensated AC line (especially a cable) can be, and it is a core reason long submarine crossings and very long bulk transfers migrate to HVDC, which has no steady-state charging current and therefore no Ferranti rise at all.
| Line length / type | Electrical length βl | Approx. no-load rise Vr/Vs − 1 | Practical note |
|---|---|---|---|
| 150 km OHL, 400 kV | ~0.18 rad (10.3°) | ~1.6% | Usually tolerable |
| 300 km OHL, 400 kV | ~0.36 rad (20.6°) | ~6.5% | Often needs a reactor |
| 500 km OHL, 400 kV | ~0.60 rad (34.4°) | ~19% | Shunt reactor mandatory |
| 100 km EHV cable, 400 kV | ~0.30 rad (17°) | ~4.5% | Cable C is ~20–30× OHL |
| Quarter-wave (~1500 km, 50 Hz) | π/2 (90°) | → ∞ (resonance) | Theoretical limit, cos(βl)→0 |
Frequently asked questions
Why does voltage rise instead of drop on an unloaded line?
With no load, the only current flowing is the capacitive charging current drawn by the line's shunt capacitance, which leads the voltage by 90°. When this leading current passes through the line's series inductance, the resulting inductive voltage drop adds to the source voltage rather than subtracting from it. The net result is a receiving-end voltage higher than the sending end.
What is the governing equation for the Ferranti effect?
For a lossless line, Vs = Vr·cos(βl), so Vr/Vs = 1/cos(βl), where β = ω√(LC) is the phase constant and l is line length. A useful small-angle approximation is ΔV ≈ ½·ω²·L·C·l²·Vr, showing the rise grows with the square of both line length and frequency.
Why is the effect much worse on cables than overhead lines?
Underground and submarine cables have shunt capacitance per km roughly 20–40 times that of overhead lines because the conductor and sheath are separated by only a thin, high-permittivity insulation layer. Since the charging current and the Ferranti rise scale with C, cables reach dangerous voltage rise over far shorter distances, which strongly limits practical AC cable length.
How do shunt reactors fix it?
A shunt reactor is an inductor connected line-to-neutral that draws a lagging (inductive) current, directly opposing the capacitive charging current of the line. This absorbs the excess reactive power (Qc = ωC·l·V²) that causes the rise. Reactors are typically sized to compensate 60–80% of the line charging MVAR and may be switched out under heavy load.
How does the Ferranti effect relate to surge impedance loading?
At surge impedance loading (SIL = V²/Zc, with Zc = √(L/C)), the line's absorbed inductive VARs exactly balance its generated capacitive VARs and the voltage profile is flat. Below SIL (light load) the surplus capacitive VARs produce the Ferranti rise; above SIL (heavy load) the deficit produces voltage sag. Ferranti is simply the below-SIL regime.
Does HVDC suffer from the Ferranti effect?
No. The Ferranti effect is an AC phenomenon driven by 50/60 Hz charging current through shunt capacitance. HVDC lines carry direct current, so in steady state there is no continuous charging current and no distributed reactive power exchange. This freedom from charging current is a major reason long submarine and bulk transmission links use HVDC instead of AC.