Which conductor does the stroke hit?
The electrogeometric model (EGM) gives every conductor a capture circle of radius rc, and the ground a capture line at rg. Both grow with the stroke current. A descending leader attaches to the first one it touches. This page works on one cross-section of the tower over level ground. It compares the classical assumption of vertical leaders with leaders arriving at an angle, as in the middle-phase paper.
Leader mechanism
Why a leader goes to the first object within striking distance, and why leaders arrive at an angle. Animated, step by step.
EGM calculator
One cross-section, level ground. Quick Imax and shielding-failure rates for any thesis tower and phase.
3D Strike Lab
The whole span on real terrain: valleys, hillsides, polarity, animated leaders and storm statistics.
Results
copied ✓Capture geometry
Where the exposure width e/π comes from
The paper reduces the leader-angle result for a middle phase to one line: WS = (e/π)·sin²ψm ≈ e/π. Here is the picture behind it, on the T20 cross-section. Move the sliders.
- Capture circles. Every conductor is surrounded by a circle of radius rc. R, S and T hang one above another, d = 4.7 m apart, and rc is ten times larger, so on the outer side the three circles form an almost vertical wall at xw = xS + rc.
- Three bands. Adjacent circles intersect half-way between the conductors, so the wall is cut into bands of height d. A leader that crosses the middle band goes to S.
- Vertical leaders miss it. Coming straight down, a leader reaches the wall from above and is already inside R's circle (or the shield wire's). That is why S looks shielded.
- Angled leaders see it. A leader at angle ψ sees the band foreshortened to a width e·sin ψ, measured across its path.
- The ground cuts the band: e. The ground captures leaders from a height rg (the line G). Any part of the S band below G is taken by the ground first. What is left above G is the exposed height e = min{d, max[0, yS + d/2 − G]}.
- Average over angles: π. Converting to a width on the ground (÷ cos ψ) and weighting by g(ψ) = (2/π)cos²ψ:
W_S = ∫ e·sinψ · (2/π)cos²ψ / cosψ dψ = (2e/π)·∫₀^ψm sinψ cosψ dψ = (e/π)·sin²ψm ≈ e/π
Why two columns?
Shielding guides (IEEE 1243, CIGRE TB 63) assume leaders come straight down. Then a phase hanging directly below another is always shielded, because its circle is hidden behind the circle above. Real leaders arrive at an angle. With the distribution g(ψ) = (2/π)·cos²ψ they can reach the side of the hidden circle, up to a maximum current set by the clearance above the ground. On the thesis towers this exposes the middle phase, which a two-phase (R/T) arrester scheme leaves unprotected.
What this page leaves out
It assumes level ground. A valley or a hillside beside the line raises the clearance and Imax: in Table III of the paper, ground falling away at 10° raises it by 30 %. Sag along the span enters only as an average height, and both polarities use the same striking distance. For terrain, sag and the whole span, use the 3D Strike Lab. For what happens after a stroke reaches the phase, see the Relief Wave Lab.
Formulas used
Striking distances (I in kA, r in m)
IEEE 1243: rc = 10·I^0.65, rg = (3.6 + 1.7·ln(43 − h))·I^0.65 for h < 40 m · Armstrong–Whitehead: 6.7·I^0.8, 6.0·I^0.8
Eriksson: R = 0.67·h^0.6·I^0.74 for each conductor · simplified LPM: kL·I^0.65·(h/10)^0.15, kL = 10·500/Ei
Vertical leaders
Leader-angle distribution (paper, eq. 2)
Middle phase below the top phase over level ground: W = d/π = 1.50 m when fully exposed, and Imax = [(hS + d/2)/Ag]^(1/b). Check: hS = 40 m gives 12.42 kA (Brown–Whitehead) and 9.20 kA (Love), as in Table I.
Shielding-failure rate
NL = Ng·(28·h0.6 + b)/10 (Eriksson), b = 2·xg.