Galloping: How a Sliver of Ice Turns a Power Line Into a Standing Wave

How it works: a round cable becomes a bad airfoil

A bare stranded conductor is essentially a circular cylinder. Its lift coefficient C_L is zero at all wind angles by symmetry, and its drag coefficient C_D is roughly constant (≈ 1.0–1.2 in the subcritical range, Re = ρUD/µ ≈ 10⁴). Round cylinders can shed vortices, but they cannot gallop — there is no lift slope to destabilize them.

Add a thin, eccentric ice accretion — a wedge, D-shape, or crescent that grows preferentially on the windward or stagnation face during freezing rain, rime, or wet snow — and the cross-section loses its symmetry. Now the effective angle of attack α (set by the wind direction plus the conductor's instantaneous vertical velocity) produces a real, sloped lift curve C_L(α). The section behaves like a stalled, low-aspect airfoil with a lift coefficient that can decrease with increasing α over some window — a negative lift slope, ∂C_L/∂α < 0.

The key coupling is kinematic: when the cable moves upward with vertical velocity ẏ into a horizontal wind U, the air it feels arrives from below at a relative angle α ≈ ẏ/U (small-angle). That self-induced change in α changes the aerodynamic force. If the force change pushes the cable further in the direction it is already moving, the airflow is adding energy each cycle instead of removing it — self-excited oscillation.

The Den Hartog criterion: one sign test that predicts everything

Model one point on the span as a mass-spring-damper moving vertically: m·ÿ + c·ẏ + k·y = F_y(α), where F_y is the vertical component of the aerodynamic force. With relative angle α ≈ ẏ/U, linearizing the lift and drag gives a vertical force whose velocity-proportional part is the aerodynamic damping:

c_aero = ½ ρ U D · ( ∂C_L/∂α + C_D )

This term adds to the structural damping c. The system is unstable when the total damping goes negative, i.e. when the aerodynamic part is negative enough to overwhelm c. Since ½ρUD and C_D are positive, the condition for galloping is the celebrated Den Hartog criterion:

• ∂C_L/∂α + C_D < 0 — galloping possible.
• Physically: the lift must fall with angle of attack faster than drag holds it back. A steep negative lift slope (a section near stall) is what kills the damping.

Because C_D ≈ 1 for iced cables, you need ∂C_L/∂α more negative than about −1 rad⁻¹. A clean cylinder has ∂C_L/∂α = 0, so it is unconditionally stable. Certain ice shapes — the classic D-section and thin crescent — give ∂C_L/∂α ≈ −2 to −4 rad⁻¹ over a band of angles, easily triggering instability. This is a 1-degree-of-freedom flutter: unlike bridge-deck flutter it needs no torsion-plunge coupling, which is why a single span can gallop at a single mode.

The physics: negative damping, energy per cycle, and limit cycles

The linear criterion tells you whether motion grows; it does not tell you how big it gets. Growth is bounded because C_L(α) is nonlinear — as amplitude increases the section swings through angles where ∂C_L/∂α + C_D turns positive again, restoring damping. The oscillation settles into a limit cycle: the net energy fed in over one cycle equals the energy dissipated by structural and residual aerodynamic damping.

Energy input per cycle scales as the closed integral ∮ F_y dy around the loop; a fat negative-damping region gives a large limit-cycle amplitude. Typical galloping amplitudes reach a substantial fraction of the conductor sag — for a 300 m span sagging 8 m, peak-to-peak motion of 3–8 m is routine, and full-sag galloping that lets adjacent phases touch has been recorded.

The unstable modes are the low transverse and vertical sway modes of the catenary, with natural frequencies f_n = (n/2L)·√(H/m'), where H is the horizontal tension, m' the mass per unit length, and L the span. For H ≈ 30 kN, m' ≈ 1.6 kg/m, L = 300 m, the first mode is ≈ 0.24 Hz — squarely in the galloping band. Galloping most often appears as one or two vertical loops per span, sometimes with a horizontal figure-eight when torsional coupling twists the ice into a fresh angle each cycle.

Controlling variables and the design trade-offs

Whether a given line gallops depends on a short list of parameters, and mitigation is the art of nudging them:

• Ice shape and mass (the trigger): engineers cannot stop ice, but the ice's rotational position sets α. Increasing the conductor's torsional stiffness keeps the ice from rotating into the worst angle. Bundle conductors (2–4 subconductors on spacers) are torsionally stiff and gallop differently — and more predictably — than single conductors.
• Tension H and mode frequency: raising tension raises f_n and shifts sag, but higher tension worsens fatigue and aeolian vibration — a direct trade-off. You cannot simply out-tension galloping.
• Structural damping ζ: conductors have tiny damping (ζ ≈ 0.001–0.01). Because c_aero can be strongly negative, adding modest damping rarely suppresses fully-developed galloping — you must instead attack the aerodynamics or detune the modes.
• Span geometry: long, slack spans with low fundamental frequency are the worst. Phase-to-phase and phase-to-ground clearances set how much amplitude is tolerable before a flashover; adding electrical clearance is expensive tower steel.

The central design tension: the same round, low-tension conductor that is cheap, low-loss, and quiet in aeolian terms is exactly the one that galloples badly once iced. Mitigation adds hardware, weight, and drag rather than changing the conductor itself.

Sizing and quantitative scale: from wind tunnel to span

A quantitative galloping assessment runs roughly like this:

• Step 1 — get the aerodynamic coefficients. Wind-tunnel test the iced (or simulated D-shape) section over α = 0–360° at the field Reynolds number (Re ≈ 2×10⁴–10⁵) to obtain C_L(α), C_D(α), and the moment C_M(α). Compute ∂C_L/∂α + C_D and flag the unstable angle bands.
• Step 2 — reduced velocity check. Galloping onset scales with reduced wind speed U_r = U/(f·D). Onset is typically U_r ≈ 20–60; below the critical U_r the negative damping cannot overcome structural damping. Design wind of 7–15 m/s transverse is where most galloping occurs.
• Step 3 — mass-damping (Scruton) number. The lumped parameter Sc = 2 m ζ / (ρ D²) sets how much aerodynamic negative damping the span can absorb before going unstable; low Sc (light, low-damping cable) means low galloping threshold.
• Step 4 — amplitude and clearance. Estimate the limit-cycle amplitude from the nonlinear C_L(α) loop or empirical galloping ellipses (single conductor: major axis ≈ 0.6–1.0 × sag). Check that the swept ellipse preserves minimum air gaps — e.g. ~2.6 m phase-to-phase at 230 kV.

For a 795 kcmil ACSR 'Drake' conductor (D ≈ 28 mm, m' ≈ 1.63 kg/m) on a 300 m span at 20% RTS tension (~30 kN), f₁ ≈ 0.24 Hz, and a moderate crosswind can drive metre-scale galloping ellipses — the numbers that motivate every anti-galloping device on the market.

Real hardware and mitigation: detuners, dampers, and spacers

Because you cannot economically stop the ice, mitigation targets the mechanism — either detuning the torsion so the ice can't reach the unstable angle, or disrupting the coherent aerodynamic forcing along the span:

• Torsional detuners / anti-galloping devices (AGDs): eccentric pendulum weights (e.g. Richardson, TDD detuners) clamped along the span change the torsional-to-vertical frequency ratio, so ice-induced twist decouples from the vertical sway and the Den Hartog feedback breaks.
• Interphase spacers: insulating rods installed between phases mechanically prevent conductors from clashing even if they gallop — they don't stop motion, they stop the flashover. Cheap and widely used as a last line of defense.
• Air-flow spoilers and aerodynamic devices: loose-hanging 'air-flow spoilers' and twisted-pair conductors vary the effective angle along the span so no long segment sits at one unstable α; they trade a little steady drag for lost coherence.
• Spacer-dampers on bundles: bundle spacers add torsional stiffness and damping simultaneously and are standard on EHV lines.
• Bundle geometry and de-icing: some utilities run higher current to melt ice or use mechanical de-icers on critical spans, though these are operational, not structural, fixes.

Standards-wise, CIGRE Technical Brochure 322 ('State of the Art of Conductor Galloping') is the field reference; ASCE Manual 74 and IEC 60826 cover overhead-line loading and reliability including ice-plus-wind combinations that drive galloping design.

Failure modes, limits, and best practice

Galloping rarely fails a line in a single cycle; it fails it three ways over time and over storms:

• Electrical flashover: the dominant nuisance. Phases swing close enough that the air gap breaks down, tripping the line — repeated auto-reclose attempts into a still-galloping span cause outages and can burn conductors.
• Fatigue and hardware failure: the low-frequency, high-amplitude bending at suspension clamps and tie points accumulates cycles. Over a multi-hour galloping event a span can log 10³–10⁴ large cycles, cracking conductor strands, tie wires, armor rods, and insulator hardware. Tension fluctuations of ±10–30% pull on tower cross-arms and can bend or buckle them.
• Cascading tower loads: synchronized galloping across a line section applies large dynamic longitudinal and vertical loads that were never in the static design envelope, risking cross-arm or even structure collapse — the reason ice-plus-wind is a governing load case.

Best practice: screen every new line's sections against the Den Hartog criterion using region-specific ice/wind statistics; design clearances for a plausible galloping ellipse rather than the static conductor; fit detuners or interphase spacers on historically galloping spans; and log outages to build a gallop map. The limit of prediction is real: ice shape is stochastic, tunnel coefficients scatter, and a span may gallop one storm and not the next at the same wind — which is why mitigation is layered (detune + limit-motion + limit-consequence) rather than betting on a single device.

Galloping versus vortex-induced (aeolian) vibration on overhead conductors

Property	Galloping	Aeolian vibration (vortex shedding)
Frequency	0.1–1 Hz (1st–3rd sway mode)	5–150 Hz (Strouhal, St ≈ 0.2)
Amplitude	0.3–10 m (order of the sag)	0.01–1 conductor diameter
Trigger	Asymmetric ice + moderate crosswind	Steady low wind 1–7 m/s, bare conductor
Mechanism	1-DOF flutter, negative aero-damping	Kármán vortex lock-in resonance
Instability test	∂C_L/∂α + C_D < 0	f_s = St·U/D near a natural mode
Mitigation	Anti-galloping detuners, interphase spacers	Stockbridge dampers, spacer-dampers