Lightning protection of a 150 kV transmission line

From a thundercloud to a tripped breaker

A lightning outage is a chain of events. Charge builds in a cloud, a leader finds the line, a current pulse becomes travelling waves, and the insulation either holds or flashes over. The earthing and the arresters decide which. This lab follows the chain on towers T16–T25 of the SBSLM–SGKIL double-circuit line, step by step, with a working model for each link. Use the menu on the left to jump to any page.

+ + +− − − −+ 1 2 3 4 5 6 7 8 9 chargeleaderattachmentcurrent wavesinsulationearthingarresterdecision

The four layers

1
Interception — two earth wires (GSW 55 mm², ±3.2 m) shield the phases. Failure mode: a shielding failure, which oblique leaders make possible on the middle phase. EGM · 3D
2
Down path — tower body (ZT ≈ 184 Ω). MDG adds four A3C 70 mm² down conductors bonded to the body.
3
Dissipation — the footing earthing. MDG adds a rod cluster in parallel with the existing MRG. The footing resistance sets the footing voltage and drives backflashover. Backflashover Lab
4
Insulation & arresters — 12 × U120B strings with arcing horns (CFO ≈ 884 kV); EGLA arresters on phases R and T. Relief wave

Main result for the thesis tower

T20 · Bb+15 · 46 m · ρ = 797 Ω·m · model R0 = 66 Ω

The guided path, in nine steps

Read in order, or click a number on the diagram. Each step links to the pages where you can try it.

1
The cloud

Charge builds up in the thundercloud

Updrafts separate charge. In a typical thundercloud the main negative region sits in the middle, positive charge sits above it, and a small positive pocket sits below. The ground underneath collects charge of the opposite sign. When the field at the cloud base is strong enough, a leader starts towards the ground. About nine flashes in ten lower negative charge. The rest are positive and come from the upper region or the anvil.

12.8 flashes/km²/yr at T20–T21~4 line average9 : 1 negative : positive
2
The leader

A stepped leader comes down, and the current sets its reach

The leader descends in steps of tens of metres. It does not know where the line is until it is close. The electrogeometric model says that a leader is captured once it comes within a striking distance of a conductor, and that distance grows with the current the stroke will carry. A weak stroke can come close before being captured. A strong one is caught far out.

rc = 7.1·I0.7540 m at 10 kA93 m at 31 kA170 m at 68.6 kA
3
Attachment

Shield wire, phase or ground: who wins?

The two shield wires are there to win. If they do, the stroke is shielded, and the risk becomes backflashover (step 7). If a phase wins, it is a shielding failure. Classical design assumes vertical leaders, and then a middle phase directly below the top phase looks perfectly shielded. Real leaders arrive at an angle. Over a valley or on a tall tower the middle phase is exposed to strokes of up to 13–17 kA. On the T20–T21 span that is enough to flash it over.

Vertical leaders: S shieldedOblique leaders: WS = d/π = 1.5 mImax over the valley: 13–17 kA
4
The return stroke

The current pulse: peak, front and tail

When the leader attaches, a return stroke drives a current pulse through the struck object. Its peak decides where it struck (step 2). Its front decides how fast the insulation is stressed. Its tail decides how much charge, and therefore energy, an arrester must absorb. Median first-stroke peaks are 31 kA in temperate data and 22–24 kA in Indonesian measurements.

Median 31 kA CIGRE24 kA Indonesia22 kA this line (LLS)Steepness ≈ 25 kA/µs tropical
5
Travelling waves

The current becomes voltage waves on the line

A conductor has a surge impedance of a few hundred ohms, so a current pulse injected into it becomes a voltage wave. A stroke on a phase splits in two and runs towards both towers at nearly the speed of light. A stroke on the shield wire runs down the tower, reflects at the footing and couples onto the phases. Every tower, flashover and arrester sends reflections back along the line.

−10 kA on phase S ≈ −1.6 MVc ≈ 300 m/µs323 m to each tower: 1.08 µs
6
Insulation

Does the insulator string flash over?

Each string has 12 glass discs with arcing horns. Flashover is not instant. Above a critical field a leader grows across the horn gap, and it needs time to bridge it. A short, high pulse can pass without flashover, while a lower but longer one may not. The few tenths of a microsecond this takes are also the time an arrester has to act.

CFO ≈ 884 kV (spec)EMTP calibration 912.5 kV on a 1.44 m gapTime to flashover ≈ 0.7–0.9 µs
7
Earthing

The footing decides the tower voltage

When the shield wire is struck, the current flows down the tower into the soil. A high footing resistance lifts the whole tower, and the insulators see the difference between the tower and the phase voltage. If that is too large, the insulator flashes from the tower side: a backflashover. Soil ionisation lowers the resistance at high currents. Multi Direct Grounding, a direct grounding system, top grounding, counterpoise, multi-rod grounding and two buried mesh grids are seven ways to lower it further.

T20: R₀ = 66 Ω, ρ ≈ 800 Ω·mRi = R₀/√(1 + I/Ig)Footing range on T16–T25: 7–66 Ω
8
Arresters

Line arresters: where, and how many?

A gapped line arrester (EGLA) clamps its phase to about 300 kV once its gap sparks over. It also sends a relief wave back along the line. The usual two-phase scheme protects the top and bottom phases against backflashover and leaves the middle phase bare. A stroke on the middle phase then flashes it, unless that phase has arresters too. A single arrester reaches only 60–90 m along the line, so a long span needs one at each end.

Clamp: 306 kV at 10 kARelief wave crosses 645.6 m in 2.15 µsOne-end reach xp ≈ 60–90 mBoth ends: 19 / 19 cases protected
9
Decisions

From one stroke to a decision for the whole line

One stroke is an event. A protection decision needs rates: how often each mechanism trips the line, which spans and phases carry the risk, and what each option costs per flashover avoided. After an outage, the burned horns, the arrester counters and the lightning-location record can tell which mechanism it was. For the 2 May 2026 outage they point to a direct strike on the middle phase, not a backflashover.

49 lightning faults on 36.5 km (2022–2025)≈ 34 per 100 km per yearMiddle phase per valley span side: 1 per 1,100–4,700 yr

Thesis tower window

Click a row to load the tower into every module. "Assigned" footings are soil-class placeholders, not measurements.

TowerTypeH (m)Base (m)R₀ (Ω)ρ (Ω·m)Ig (kA)Ng localSoilFooting data
Grounding · MDG · 1 Design

Multi Direct Grounding — lattice tower 150 kV

Four A3C 70 mm² down conductors bonded to the body, and a 4-rod cluster under the tower in parallel with the existing earthing. Redrawn from the PLN UIP3B Sumatera drawing on the selected thesis tower. Click any part for its specification.

Elevation

red = MDG additions

Select a component

Click the earth wire, PG clamps, down conductors, cable ties, busbar, rods, buried connectors or the existing leg earthing.

Base plan — MDG rods in parallel with the existing MRG

Bill of materials

ItemSpecification (drawing)Qty
Down conductorA3C (AAAC) 70 mm²4 runs
PG clamp55–70 mm² (to GSW), 70–70 mm²per run
Cable tieStainless steel, 30 cmeach panel
TerminalSkun 70 mm (bolt + nut), busbar 1 m, 0.2 m above ground1
Earth rodAngle steel 50 × 50 × 4 mm, 2 m (1.8 m buried)4
Buried connectorGSW 70 mm², ≥ 0.5 m deep4
Busbar supportAngle 50 × 50 × 4, M10 bolted joint2

Design review

Numbers follow the selected tower and the inputs on the Earthing and Lightning pages.

    Grounding · MDG · 2 Earthing

    Rod cluster in parallel with the existing earthing

    Method of moments in uniform soil. Each conductor is split into segments with uniform leakage, bonded at one potential, with an image for the soil surface. The existing earthing is calibrated so the model reproduces the measured R₀.

    Surface potential (% of GPR)

    contours every 10 %

    Combined R vs rod length

    Combined R vs square side

    at the current rod length

    Impulse resistance with soil ionisation

    Ri = R₀ / √(1 + I/Ig)

    Target finder

    Shortest rod length that brings the combined footing to the target, keeping the other inputs.

    Grounding · MDG · 3 Lightning

    Does the MDG change the backflashover rate?

    A negative stroke to the tower top. Lossless travelling-wave model: tower (ZT, 0.8c), two earth-wire spans to the neighbouring towers, current-dependent footing. Insulator voltage = crossarm voltage − coupled voltage − power-frequency voltage, compared with the CIGRE short-tail strength.

    Worst-phase insulator voltage

    Tower-top and footing voltage

    lossless lines: reflections show as sharp steps

    Critical current vs footing resistance

    this tower, this stroke front

    Probability a first stroke exceeds I

    Down conductors and tower surge impedance

    Four A3C 70 mm² runs inside the lattice body, bonded at every panel.

    Grounding · DGS · 1 Design

    Direct Grounding System — isolated earth wire and separate rods

    A second GSW 70 mm² is laid over the peak on post insulators (isolator tumpu), with its ends bent upward, and joined to the existing earth wire by parallel-groove clamps. An A3C-S 70 mm² lead runs down on standoff insulators, through a lightning counter and a PVC sleeve, to a rod string set apart from the tower (lepas tower). Click any part for its specification.

    Elevation

    blue = DGS earth wire · red = insulated lead

    Select a component

    Click the DGS earth wire, post insulators, PG clamps, insulated lead, counter, PVC sleeve or rods.

    Peak — top view

    Base plan — DGS rods apart from the tower earthing

    Bill of materials

    ItemSpecification (drawing)Qty
    DGS earth wireGSW 70 mm², loop over the peak, ends bent up1 set
    Post insulatorIsolator tumpu, 20 kV class, on steel supports10 + standoffs
    PG clamp55–70 mm² to existing EW; 70–70 mm² joints2 + 2
    Down leadA3C-S 70 mm², not touching the body1 run
    Lightning counterEP 67, fixed on the lead1
    SleevePVC 3/4″, 80 cm at grade1
    Earth rodRod Ø 5/8″, 2 m, top at 1 m; moulding welding4
    PlateAluminium 500 × 500 × 2 mm, skun and bolt1

    Design review

    Numbers follow the selected tower and the inputs on the Earthing and Lightning pages.

      Grounding · DGS · 2 Earthing

      Separate rod string and its coupling to the tower earthing

      The DGS rods and the tower earthing are two electrodes in the same soil. The two-port model gives each self-resistance and the mutual resistance between them, from the same method-of-moments solver as the MDG.

      Surface potential (% of GPR)

      after the standoff flashes (bonded)

      DGS rod resistance vs number of rods

      Grounding · DGS · 3 Lightning

      Does the isolation survive a stroke?

      The insulated lead and the tower are modelled as two coupled vertical conductors. A switch across the standoff insulators closes when the voltage between lead and tower exceeds their breakdown value. After that the DGS acts as a bonded parallel earth.

      Voltage across the standoff (lead − tower)

      Current into the two earthings

      Worst-phase insulator voltage

      Critical current and backflashover rate

      Same stroke front, insulation and exposure on all three options. No arresters.

      Grounding · Counterpoise · 1 Design

      Counterpoise — buried radials and end rods only

      Nothing is added on the tower. GSW 55 mm² runs from two diagonal legs (skun 70, M12 bolt) down to 1 m depth, then 20 m along the line, ending in 2 m rods joined by moulding welding and kuku macan clamps with an aluminium plate. Buried GSW 55 mm² bonds join the legs on each side. Click any part for its specification.

      Plan

      red = radials · blue = leg bonds

      Select a component

      Click a radial, the end rods, the leg bonds or the tower legs.

      Section

      Bill of materials

      Design review

      Numbers follow the selected tower and the inputs on the Earthing and Lightning pages.

        Grounding · Counterpoise · 2 Earthing

        Radials outside the leg footprint

        The radials are bonded to the existing earthing and solved together with it by the method of moments. During a stroke front only the first part of each radial acts. The impulse-effective resistance uses Grcev’s effective length.

        Surface potential (% of GPR)

        R vs radial length

        solid = low-frequency · dashed = impulse
        Grounding · Counterpoise · 3 Lightning

        Does the counterpoise change the backflashover rate?

        The same travelling-wave model as the MDG page, with the footing set to the impulse-effective resistance of the counterpoise. The table at the bottom puts all seven grounding options side by side on the same tower and stroke.

        Worst-phase insulator voltage

        Critical current vs footing resistance

        this tower, this stroke front

        All grounding options — same tower, same stroke

        Each option uses its own page's inputs. No arresters. For arresters and costs, use the Compare Lab.

        Grounding · Mesh grid type B · 1 Design

        Mesh grid type B — a 6 m square grid under the tower

        PLN Desain B: GSW 55 mm² buried 1 m in a 4 × 4 grid at 2 m spacing (6 m × 6 m), a 2 m copper-bonded 5/8″ rod at every node joined by exothermic (Cadweld) moulding, and the four corner nodes run diagonally to the four legs, fixed to the tower bracing with a tower earth clamp. Click any part for its specification.

        Plan

        red = mesh · blue = leg links · ◎ = rods

        Select a component

        Click the mesh, a rod, a leg link or the tower legs.

        Section

        Bill of materials

        Design review

        Numbers follow the selected tower and the inputs on the Earthing and Lightning pages.

          Grounding · Mesh grid type B · 2 Earthing

          A compact grid inside or around the leg footprint

          The grid is bonded to the existing earthing and solved together with it by the method of moments: every mesh conductor, rod and leg link is a set of segments, and mutual resistance is included. A compact grid sits in soil the legs already use, so the gain is less than the grid alone would suggest.

          Surface potential (% of GPR)

          R vs mesh spacing

          both mesh layouts, this tower
          Grounding · Mesh grid type B · 3 Lightning

          Does the mesh grid change the backflashover rate?

          The same travelling-wave model as the other grounding pages, with the footing set to the impulse-effective resistance of the tower plus grid. The table at the bottom puts all seven grounding options side by side on the same tower and stroke.

          Worst-phase insulator voltage

          Critical current vs footing resistance

          this tower, this stroke front

          All grounding options — same tower, same stroke

          Each option uses its own page's inputs. No arresters. For arresters and costs, use the Compare Lab.

          Grounding · Mesh grid type B (modified) · 1 Design

          Mesh grid type B, modified — a cross with pointed arms

          The approval drawing (TPG): the same 16 rods and GSW 55 mm² at 1 m, laid out as a cross of five 2 m cells with a pointed tip (2.24 m diagonals) on each arm, reaching 5 m from the centre along and across the line. Each leg connects to the nearest arm node with skun GSW 55–70 and an M12 × 35 bolt on the stub. Rods at least 2 m apart, at least 2 m long. Click any part for its specification.

          Plan

          red = mesh · blue = leg links · ◎ = rods

          Select a component

          Click the mesh, a rod, a leg link or the tower legs.

          Section

          Bill of materials

          Design review

          Numbers follow the selected tower and the inputs on the Earthing and Lightning pages.

            Grounding · Mesh grid type B (modified) · 2 Earthing

            A compact grid inside or around the leg footprint

            The grid is bonded to the existing earthing and solved together with it by the method of moments: every mesh conductor, rod and leg link is a set of segments, and mutual resistance is included. A compact grid sits in soil the legs already use, so the gain is less than the grid alone would suggest.

            Surface potential (% of GPR)

            R vs mesh spacing

            both mesh layouts, this tower
            Grounding · Mesh grid type B (modified) · 3 Lightning

            Does the mesh grid change the backflashover rate?

            The same travelling-wave model as the other grounding pages, with the footing set to the impulse-effective resistance of the tower plus grid. The table at the bottom puts all seven grounding options side by side on the same tower and stroke.

            Worst-phase insulator voltage

            Critical current vs footing resistance

            this tower, this stroke front

            All grounding options — same tower, same stroke

            Each option uses its own page's inputs. No arresters. For arresters and costs, use the Compare Lab.

            Grounding · Top grounding · 1 Design

            Top grounding — a lead from the peak to its own deep electrode

            An AAACS 70 mm² conductor bridges the two earth wires over the peak (PG clamps 70–55 and 70–70), carried on post insulators (isolator tumpu) on a UNP channel. The lead hangs from a suspension insulator with a U bolt, strain clamp, socket and ball clevis, and runs down inside the tower body. At the base it passes through a 60 × 30 cm concrete pad and an angle-iron frame (wire clamp, skun 70), then an Al/Cu joint sleeve 70/95 and BC 70 mm² to a 1.06 m counterpoise bar 3 m below grade. Click any part for its specification.

            Peak

            pink = AAACS 70 mm² · yellow = PG clamps

            Base detail

            Select a component

            Click the bridge, the clamps, the insulators, the lead, the pad or the bar.

            Bill of materials

            Design review

            Numbers follow the selected tower and the inputs on the Earthing and Lightning pages.

              Grounding · Top grounding · 2 Earthing

              One short, deep electrode under the tower

              The bare lead below the pad and the counterpoise bar form a separate electrode. It is solved together with the existing earthing as a two-port by the method of moments: its own resistance, the tower's, and the mutual resistance through the shared soil.

              Surface potential, bonded (% of GPR)

              Electrode R vs depth

              this soil, this bar length
              Grounding · Top grounding · 3 Lightning

              Does the top-grounding lead stay isolated?

              The same coupled lead-and-tower model as the DGS page: the lead and the tower are two coupled surge paths, the insulators at the peak are a switch that closes at Vbd, and the two electrodes share soil through the mutual resistance.

              Voltage across the peak insulators

              Worst-phase insulator voltage

              Critical current

              All grounding options — same tower, same stroke

              Each option uses its own page's inputs. No arresters. For arresters and costs, use the Compare Lab.

              Grounding · Multi Rod Grounding · 1 Design

              Multi Rod Grounding — a fan of rods at every leg

              At each leg an angle-iron connector (60 × 60 × 5, M10 × 50 bolts) is bolted to the existing stub and takes four GSW skuns. Three GSW 55 mm² radials run outward 1 m deep, 30° apart, each ending in a copper-dipped rod at least 2 m long, joined by exothermic welding. Buried GSW 55 mm² bonds run from each leg towards the other legs. Click any part for its specification.

              Plan

              red = radials · blue = leg bonds · ◎ = rods

              Select a component

              Click a radial, a rod, a leg bond, a connector or the tower legs.

              Section

              Bill of materials

              Design review

              Numbers follow the selected tower and the inputs on the Earthing and Lightning pages.

                Grounding · Multi Rod Grounding · 2 Earthing

                Rods pushed out from every leg

                The radials, rods and leg bonds are bonded to the existing earthing and solved with it by the method of moments. Each fan starts at a leg and points outward, so the rods sit in soil the footings use less.

                Surface potential (% of GPR)

                R vs rods per leg

                this tower
                Grounding · Multi Rod Grounding · 3 Lightning

                Does MRG change the backflashover rate?

                The same travelling-wave model as the other grounding pages, with the footing set to the impulse-effective resistance of the tower plus MRG.

                Worst-phase insulator voltage

                Critical current vs footing resistance

                this tower, this stroke front

                All grounding options — same tower, same stroke

                Each option uses its own page's inputs. No arresters. For arresters and costs, use the Compare Lab.

                Method, data and limits

                How the numbers are made

                1. Earthing — method of moments

                Each conductor (leg footing, counterpoise, rod, buried connector) is split into straight segments carrying a uniform leakage current Ij. The potential at the surface of segment i is

                V_i = Σ_j (ρ I_j / 4π) [ K(p_i, s_j) + K(p_i, s_j') ] K(p, s) = ln[(r_A + r_B + ℓ) / (r_A + r_B − ℓ)] / ℓ

                s′ is the image of s in the soil surface. With all segments bonded at 1 V, R = 1 / Σ Ij. A single 3 m rod gives 33.43 Ω against 33.49 Ω from Dwight's formula R = ρ/(2πL)·[ln(4L/a) − 1].

                2. Existing earthing calibration

                The as-built MRG/counterpoise layout is not in the drawing. It is represented by four leg footings (r = 0.3 m) of equivalent depth d. If d = 3 m is not enough, four diagonal counterpoises of length Lc at 0.5 m are added. d or Lc is chosen by bisection so the model returns the measured R₀. The MDG cluster is then added to the same equipotential system, so mutual resistance is included automatically:

                R_c = (R_ex R_mdg − R_m²) / (R_ex + R_mdg − 2 R_m)

                3. Soil ionisation (CIGRE TB 63)

                R_i = R₀ / √(1 + I/I_g), I_g = E₀ ρ / (2π R₀²), E₀ = 400 kV/m

                This is the footing law of the thesis EMTP model. A lower R₀ raises Ig, so ionisation helps a high-resistance footing more. This narrows the difference in impulse resistance that MDG makes.

                4. Backflashover

                Bergeron lossless lines, Δt = 5–10 ns: tower ZT = 183.84 Ω at 2.4 × 10⁸ m/s. Each earth-wire span (Zg from two GSWs at ±3.2 m, 0.96 cm diameter) is terminated at the adjacent tower in R₀,adj ∥ Zg. Crossarm voltage is interpolated linearly between top and foot. Coupling factors K come from Maxwell potential coefficients (phases at ±4.5 m, H − 2.9/7.6/12.3 m). Strength:

                CFO_NS = (0.977 + 2.82/τ) CFO, τ = (Z_g / R_i) T_s

                Ic is found by bisection on the peak current. Rates:

                N_L = N_g (28 h^0.6 + b) / 10 [strokes /100 km /yr] BFR = 0.6 N_L P(I > I_c)

                5. DGS — coupled lead and tower

                The insulated lead (Zd, single 70 mm² wire) and the tower (ZT) form a two-conductor lossless line with mutual surge impedance Zm = k·min(Zd, ZT). It is solved with the matrix form of Bergeron's method, i = Z⁻¹v − Ih. At the top a switch represents the standoff insulators. It closes when |Vlead − Vtower| ≥ Vbd. At the bottom, the DGS rods and the tower earthing are a two-port resistance matrix from the same MoM solver, with mutual resistance and ionisation of each self-term.

                [R] = [G]⁻¹, G_ab = Σ I_j(group a) for 1 V on group b

                6. Counterpoise — effective length

                A long horizontal electrode does not act over its whole length during the stroke front. The effective length used here (Grcev, 2009) is

                l_eff = 0.6 (ρ T₁)^0.43 [m; Ω·m; µs]

                The "impulse-effective" resistance is solved with each radial cut to leff and the end rods removed. That value goes into the backflashover calculation, which is the conservative choice.

                Data used (thesis model, SBSLM–SGKIL 150 kV)

                QuantityValueSource
                Tower windowT16–T25, spans 223–646 mWindow 1 EMTP spec
                Heights37 / 40 / 46 m per towerTower schedule
                Footing R₀, ρper tower; 6 measured, 4 assignedTower Resistance (CT ring)
                T20 re-measurement15.75 Ω (June 2026)Field
                ZT, vT183.84 Ω, 0.8cEMTP netlist
                GSW / phaseGSW 55 mm² ±3.2 m; Hawk ±4.5 mImplemented geometry
                CFO884 kV (0.80 × 1105, arcing horns)Insulator spec
                Nglocal 7.7–13.0 fl/km²/yrVaisala FALLS 2019–26
                Base widthAA 6 m, BB/CC 8 mTower outlines

                Assumptions and limits

                • Uniform soil. Rocky sites are often two-layer, and a resistive top layer makes short rods worse than shown.
                • The equivalent radius of an L50×50×4 angle is taken as 20 mm. R changes by about 5 % for 15–25 mm.
                • The existing earthing is a calibrated stand-in. Mutual resistance with the MDG depends on where the real MRG conductors run.
                • Low-frequency resistance plus ionisation only. For electrodes this short (≪ effective length) the frequency dependence is small. Long counterpoises would need a frequency-dependent model.
                • The backflashover model is a screening model (CIGRE/IEEE 1243 class). It has no leader-progression model, corona or multiple strokes, and no EGLA arresters (R/T arresters would raise Ic on those phases; the Backflashover Lab adds gapped arresters and leader progression). Use the thesis EMTP-RV model for event-level work.
                • BFR assumes every tower is like the selected one. It is a per-tower indicator, not the line rate. For reference, the line recorded 49 lightning faults in 2022–2025 on 36.5 km, about 34 per 100 km per year.

                References

                1. CIGRE TB 63, Guide to procedures for estimating the lightning performance of transmission lines, 1991; CIGRE TB 839, 2021.
                2. IEEE Std 1243-1997, Guide for improving the lightning performance of transmission lines.
                3. IEEE Std 80-2013, Guide for safety in AC substation grounding.
                4. A. R. Hileman, Insulation Coordination for Power Systems, CRC, 1999, ch. 10.
                5. E. D. Sunde, Earth Conduction Effects in Transmission Systems, Dover, 1968.
                6. H. B. Dwight, "Calculation of resistances to ground," AIEE Trans., 1936.
                7. PT PLN (Persero) UIP3B Sumatera, Desain Multi Direct Grounding Tower Lattice 150 kV, Modifikasi DGS ke MDG, and the DGS approval drawing (TPG).
                8. L. Grcev, "Impulse efficiency of ground electrodes," IEEE Trans. Power Delivery, vol. 24, no. 1, 2009.
                9. IEC 62305-3:2010, separation distance of isolated down conductors (s = ki·kc/km·l).