Lightning protection for overhead distribution lines

29th
International Conference on
Lightning Protection
23rd
– 26th
June 2008 – Uppsala, Sweden
LIGHTNING PROTECTION OF OVERHEAD POWER
DISTRIBUTION LINES
Alexandre Piantini
University of São Paulo
piantini@iee.usp.br
Av. Prof. Luciano Gualberto, 1289, 05508-010, São Paulo – SP, Brazil
Abstract - Distribution lines are often located in areas with high ground flash densities, being therefore subject to
lightning-caused power interruptions. This paper presents an overview of the various types of lightning overvoltages that can
arise on overhead power distribution networks, as well as typical voltage waveforms. The effectivenesses of the most
important methods for mitigating such overvoltages on MV and LV networks are also discussed.
1 INTRODUCTION
In recent years, the growing use of sensitive electronic devices and the increasing demand of utility customers for
stability of the power supply have stressed the importance of improving the reliability and power quality levels of
electric systems. As lightning is a major source of faults on overhead lines and damages to or malfunction of sensitive
electronic equipment, it is essential to evaluate the lightning electromagnetic environment in order to mitigate its effects
and improve the power system quality. Many studies have been carried out, especially on medium voltage (MV) lines,
aiming at obtaining a better understanding of the characteristics of the lightning overvoltages.
More recently, special attention has been drawn to the transients on low-voltage (LV) systems. As the surge
withstand capabilities of LV networks are much lower than those of MV lines, they are more susceptible to lightning-
caused disturbances. There are various ways by which lightning can disturb low-voltage lines. Transients may be
originated from direct strokes (to the MV or LV networks or to end user installations) or indirect ones (either intracloud
or cloud-to-ground flashes).
This paper presents the major mechanisms by which overvoltages are stemmed from lightning in both medium and
low-voltage overhead power distribution networks, as well as typical surge waveforms. The effectivenesses of the main
protective measures that can be utilized to improve the line lightning performance are also discussed.
2 MV NETWORKS
The basic measures that can be applied to improve the lightning performance of MV distribution lines involve the
increase of the line insulation withstand capability [1 - 4], the use of periodically grounded shield wires [1, 4 - 8], and
the installation of surge arresters along the line [1, 2, 6 - 15]. In this Section, some typical overvoltages are presented
and the impacts of the protection alternatives on the reduction of the number of line flashovers, with emphasis on the
use of shield wires and surge arresters, are discussed.
2.A Lightning Overvoltages
Lightning can produce overvoltages when it hits either the line conductors (direct strokes) or a point in the vicinity of
the distribution network (indirect strokes).
2.A.1. Direct Strokes
If a flash hits an overhead line, the current injected into the conductor is divided at the strike point, giving rise to two
voltage waves that propagate in opposite directions. The prospective magnitude of these voltages can be estimated by
multiplying the current that flows in each direction (half of the stroke current) by the characteristic impedance of the
line, which is normally in the range of 400 Ω to 500 Ω. Therefore, for a line characteristic impedance of 400 Ω and a
stroke current of 10 kA, whose probability of being exceeded is larger than 90 %, the corresponding overvoltage is
2000 kV, which is far beyond the line insulation level. As a consequence, multiple flashovers occur between the
1

conductors and also to earth in different points of the line. Customers around the fault location experience both a
voltage sag during the short-circuit and a momentary interruption when the breaker opens to clear the fault.
A typical overvoltage caused by a direct strike to a MV line is illustrated in Figure 1. The voltage, obtained from
calculations performed by Mirra et al. [16], is characterised by a few short spikes, produced by multiple line insulation
breakdowns, followed by a slower front component whose amplitude is somewhat lower than the insulation level of the
line.
Direct strikes usually do not cause permanent damages to lines with bare conductors as long as the fault duration is
limited by a short-circuit protection device. On the other hand, in the case of lines with covered conductors, the coating
prevents the footpoint of the power frequency arcing current from moving along the line, and therefore a flashover
between phases for such lines may cause a mechanical breakdown of the conductors [17].
Fig. 1 – Example of a typical overvoltage due to a direct strike to the MV line (adapted from [16]).
2.A.2. Indirect Strokes
Although the overvoltages associated with direct strikes to the line are much more severe, those induced by nearby
lightning have a higher frequency of occurrence and are usually responsible for a greater number of line flashovers and
supply interruptions on systems with rated voltage 15 kV or less. Due to the impacts of the lightning induced
overvoltages on the performance and power quality of distribution systems, several theoretical and experimental studies
have been conducted in order to better understand their characteristics or to assess the effectiveness of the methods that
can be used for their mitigation [18 - 53].
The induced voltage magnitudes and waveforms depend on many lightning parameters and are substantially affected
by the network configuration. The evaluation of such transients entails the calculation of lightning fields, which are
defined by the spatial and temporal distribution of the stroke current along the channel, as well as by the earth electrical
parameters. A suitable coupling model is required for the analysis of the electromagnetic interaction between the field
and the line conductors. In [18] Nucci and Rachidi present different approaches and formulations to describe the
coupling between an external electromagnetic field and an overhead line. The effects of a lossy earth and the case of
multiconductor lines are also dealt with, and the influences of various parameters on the induced voltages are
investigated.
With regard to experimental investigations, Yokoyama et al. presented in [45 - 47] simultaneous measurements of
induced voltages and the corresponding stroke currents, thus allowing direct comparisons between measured and
calculated results.
The appraisal of three different theories for computation of lightning induced voltages on overhead lines presented in
[48] concludes that the Rusck model [31] leads to consistent results. However, in its original form the electric field is
assumed to be constant in the region between the line and the ground, the lengths of the line and of the stroke channel
are assumed to be infinite, only straight lines can be considered and thus realistic configurations cannot be taken into
account. These restrictions limit the application of the model and an extension was proposed by Piantini and
Janiszewski [49]. The so-called Extended Rusck Model (ERM) overcomes these limitations and enables also to take
into account the incidence of lightning flashes to nearby elevated objects, the occurrence of upward leaders, and the
presences of a periodically earthed shield wire (or neutral conductor) and equipment such as transformers and surge
arresters. Lines with various sections of different directions can be considered through the evaluation of the correct
2

propagation time delays for the elementary voltage components that determine the induced voltage at a given point of
the line.
The validity of the ERM has been demonstrated from various comparisons of theoretical and experimental results [49
- 51] for the case of an electromagnetic field radiated by a lightning channel perpendicular to the earth plane. This is
indeed the hypothesis under which the Rusck model was developed and Cooray [52] and Michishita and Ishii [53] have
shown that, for this condition, it leads to results which are identical to those obtained from the more general coupling
model proposed by Agrawal et al. [54], whose adequacy has been confirmed in [7, 26, 37, 55].
The calculation of lightning induced voltages on overhead lines through the ERM is based on electric and magnetic
potentials due to the charges and currents in the channel. The inducing scalar potential associated with the charges acts
as a distributed source and is responsible for the generation of waves that propagate along the conductors. On the other
hand, the magnetic potential associated with the currents contributes with its time derivative to the total induced voltage
in each point of the line.
In the case of direct strikes to a metallic elevated object, the return stroke starts at the top of the structure. The
currents in the object and in the stroke channel are assumed to have equal magnitudes and polarities, but different
speeds and directions of propagation, as illustrated in Figure 2 for the case of a downward negative flash. The current
through the strike object propagates at a speed very close to that of light in free space, whereas in the channel the speed
is a fraction of this value.
Fig. 2 - Lightning strike to an elevated object. ht: height of the strike object; hc: height of the upper extremity of the stroke channel;
Ic and It: currents in the stroke channel and in the strike object, respectively; s1 and s1t: heights, at instant t1,
of the current front in the channel and in the strike object, respectively (adapted from [56]).
The voltages U(x,t) induced on an overhead line located in the vicinity of the strike object are obtained by adding the
component associated with the charges in the stroke channel (electrostatic component) to those associated with the
currents that propagate in the lightning channel and in the strike object (magnetic components). Thus,
),1(.
),(
.
),(
),(),(
00
dz
t
txAi
dz
t
txAi
txVtxU
h
t
h
∫∫ ∂
∂
+
∂
∂
+=
where V(x,t) is the induced scalar potential, h is the height of the line, and Ai(x,t) and Ait(x,t) are the vector potentials
associated with the currents that propagate through the stroke channel and through the strike object, respectively. The
procedure for calculating the voltage induced on an overhead line in the presence of a tall strike object, disregarding the
reflections at the bottom and top of the structure, is described in [49].
Baba and Rakov, using the finite-difference time-domain (FDTD) method, verified that the ratio between the
magnitudes of lightning induced voltages for strikes to a tall object and to flat ground increases with increasing distance
from the lightning channel (ranging from 40 m to 200 m), decreasing the current reflection coefficients at the top and at
3

the bottom of the strike object (ρtop and ρbot, respectively), and decreasing return stroke speed [33, 57]. Also, the ratio
increases with decreasing the lightning current rise time. Under realistic conditions such as ρtop = -0.5 and ρbot = 1, the
ratio is larger than unity (the tall strike object enhances the induced voltages), but it becomes smaller than unity under
some special conditions, such as ρtop = 0 and ρbot = 1.
The lightning induced voltage waveforms presented by Yokoyama et al. in [45 - 47] were measured on an
experimental line with two sections, as shown in Figure 3. They were obtained simultaneously with the stroke currents
that hit a 200 m high metallic tower situated at a distance of 200 m from the experimental line. Electrical-optical
converters were used for transmission of the obtained waveforms by optical cables and, after optical-electrical
conversion, the data were stored in magnetic tapes. Due to the characteristics of the converters, the recorded waveforms
presented a faster decay than the real ones [46]. The length of the lightning channel was assumed to be equal to 3 km
and the stroke current was assumed to propagate through the lightning channel at a constant speed of 30 % of that of
light in free space, as these parameters were not measured.
Fig. 3 - Top view of the experimental line for the measurements carried out in [45-47] (adapted from [56]).
Figure 4 presents measured and calculated voltages induced on the experimental line by downward negative flashes
that struck the tower. The corresponding stroke current waveforms are also presented. As shown in [58], the
calculations presented in Figure 4, which take into account the effects of the tower, line topology, and finite length of
the lightning channel, are in a much better agreement with the measured voltage waveforms than those performed with
the original Rusck model. Despite the differences on the wavetails, both the voltage magnitudes and wavefronts are
reasonably well reproduced by the ERM.
The observed discrepancies can be attributed partially to the representation of the stroke current waveform, which
has a significant influence on the induced voltages especially in the case of strikes to elevated metallic objects. Other
reasons for the differences are the features of the electrical-optical converters, as already mentioned, the Transmission
Line (TL) model [58] adopted for the determination of the current distribution along the lightning channel, and the fact
that no reflections were considered at the top and bottom of the structure. Greater discrepancies would probably be
observed in the case of stroke currents with steeper fronts, since in this case the current waveforms along the tower
would be affected more significantly by the reflections at the tower extremities. Finally, the assumptions of a constant
current propagation velocity of 30 % of that of light in free space (c) and of a lightning channel perpendicular to a
perfectly conducting ground plane also contribute to the deviations. For these reasons the comparisons can only be
made under a qualitative perspective. Even so, the overall agreement between theoretical and experimental results is
very reasonable and indicates the adequacy of the ERM, whose validity has also been confirmed from many other
comparisons of measured and calculated voltages, using mostly data obtained from scale model experiments.
A comparison using data obtained from the 1:50 scale model described in [26, 30, 55] is presented in Figure 5, where
the voltage and time scales are referred to the real system by applying the corresponding scale factors (1:18000 and
1:50, respectively). In this simple configuration, the line was matched at both terminations. A good agreement is
observed between the measured and calculated voltages.
4

(a)
(b)
Fig. 4 - Measured (M) and calculated (C) lightning induced voltages at the observation point indicated in Fig. 3 and corresponding
stroke current waveforms (adapted from [56]).
a) Case 81-02 [45] b) Case 86-03 [47]
(a) (b)
Fig. 5 - Calculated (C) and measured (M) induced voltages obtained from a 1:50 scale model (all the parameters
are referred to the full-scale system). Stroke current with peak value of 34 kA,
front time of 2 µs, and propagation velocity of 11 % c.
a) line topology (top view) b) induced voltages
5

In [26, 55] the experimental facility was used to validate the Agrawal et al. coupling model [54] through
comparisons with simulations performed using the LIOV-EMTP code [7, 26, 37], considering much more complex
network configurations. The LIOV-EMTP allows for the calculation of lightning induced voltages on homogeneous,
multiconductor, lossy overhead lines taking also into account the effects of downward leader electric fields and corona
[18].
The amplitudes and waveforms of the lightning induced voltages vary widely and are particularly affected by the:
- magnitude, front time and propagation velocity of the stroke current;
- distance between the line and the lightning strike point;
- occurrence of upward leaders;
- position along the line;
- soil resistivity;
- heights of the conductors;
- line configuration (horizontal or vertical, rural or urban);
- presence of a shield wire or neutral conductor;
- earth resistance;
- relative position between the stroke location and surge arresters;
- surge arrester spacing;
- surge arrester V/I characteristic.
In the case of a strike to an elevated object in the vicinity of the line, the induced voltage depends also on the height,
surge impedance and earth impedance of the structure and on the equivalent surge impedance of the lightning channel.
An example of a typical lightning induced voltage waveform, recorded by the system described in [11] on a 2.7 km
long, unenergized line matched at both terminations, is presented in Figure 6. Despite the large variation of the induced
voltage waveforms, they are usually characterised by shorter front times and shorter durations in comparison with
overvoltages caused by direct strokes.
Fig. 6 – Example of a typical lightning induced voltage.
2.B Lightning Protection of MV Lines
The frequency of flashovers associated with indirect strokes decrease substantially as the critical flashover
overvoltage (CFO) of the line structures increases, and can be virtually eliminated if the CFO is greater than 300 kV.
However, surges with higher magnitudes will then travel over long distances, increasing the stresses on line equipment.
The number of faults caused by direct strokes will remain practically unaltered unless a shield wire earthed at every
pole with low earth resistance is used or surge arresters are installed on all phases at very short intervals. A 10 m high
distribution line located in open ground in an area with ground flash density of 1/(km2
.year) collects on average
approximately 11 flashes /(100 km.year). The use of a shield wire in conjunction with arresters on every pole and every
phase is effective against direct strokes and theoretically eliminates flashovers. The shield wire protects the arresters
from excess energy dissipation, and the arresters prevent back flashovers.
In any case, a cost-benefit analysis of the possible solutions should always be performed.
6

2.B.1. Shield Wires
Although a shield wire may collect most of the flashes that otherwise would hit the phase conductors, its
effectiveness against direct strokes is very limited. The reason is that the potential rise due to the current flow through
the pole earth impedance causes a large voltage difference between the earth lead and the phase conductors, which by
its turn causes a back flashover in the great majority of the cases. Therefore, in order to mitigate the effects of direct
strikes, the shield wire should not only be earthed at every pole, but also the line should have sufficient CFO between
the earth lead and the phase conductors, and the earth resistances should be low. Earth resistances must be less than
10 Ω if the CFO is less than 200 kV, whereas for CFO in the range of 300 kV to 350 kV, a ground resistance of 40 Ω
will provide similar performance [1].
On the other hand, a shield wire may reduce the magnitudes of the overvoltages associated with nearby strokes. As
this effect is due to the coupling between the shield and phases wires, the voltage reduction will occur regardless of the
position of the shield wire with respect to the phase conductors. The greater the coupling, the more significant the
voltage reduction. The effectiveness of shield wires in improving the indirect lightning performance of distribution
lines has been investigated in [5-7, 31, 60-63].
In [5], the mitigation effect of the neutral conductor / shield wire was studied experimentally through the use of a
reduced scale model. Several tests were performed in order to investigate the influence, on the induced voltages, of
parameters such as the height of the shield wire, the earth resistance, the distance between adjacent earthing points, and
the relative position of the stroke channel with respect to the earthing points. The voltages induced on two distribution
line models, representing 10 m high, 1.4 km long lines – one with and the other without a shield wire -, both located at
a distance of 70 m from the ‘lightning’ channel, were recorded simultaneously. The voltage measurements were
performed at the points closest to the stroke location. The ‘stroke’ current waveform was representative of first
downward negative strokes (the equivalent front time Td30-90 [59], referred to the full-scale system, was about 3.5 µs),
and the earth was simulated by interconnected aluminum plates, thus representing a perfectly conducting plane. The
current propagation velocity (v) along the return stroke model was 11 % of that of light in free space.
As mentioned above, the greater the coupling between the shield wire and the phase conductors and the larger the
voltage induced on the former, the greater its influence in terms of reduction of the voltage amplitudes. This is
illustrated in Figure 7, which compares the phase-to-earth voltages induced on lines with shield wire heights of 11 m
and 7 m with that induced on a line without shield wire or neutral conductor.
Fig. 7 – Induced voltages (phase-to-earth) at the point closest to the stroke location – influence of the height of the shield wire
(adapted from [5]). Lightning current: amplitude 24.9 kA, equivalent front time 3.5 µs; stroke location in front of an earthing point,
at a distance of 70 m from the distribution line; earthing spacing: 450 m; earth resistance: 50 Ω; perfectly conducting earth.
The influence of the earth resistance (Rg) is illustrated in Figure 8, which presents the induced voltages (phase-to-
earth) corresponding to Rg = 0 Ω and Rg = 1000 Ω for the case of stroke location equidistant from two earthing points.
The lower the value of the earth resistance, the lower the induced voltage magnitude, especially if the lightning channel
7

is in front of an earthing point. However, this influence tends to decrease as the distance between adjacent earthing
points increases, the stroke current front time becomes shorter or the distance between the stroke location and the
closest earthing point increases.
Fig. 8 – Induced voltages (phase-to-earth) at the point closest to the stroke location – influence of the earth resistance (adapted
from [5]). Lightning current: amplitude 24.9 kA, equivalent front time 3.5 µs; stroke location equidistant from
two earthing points, at a distance of 70 m from the distribution line; distance of 450 m between
adjacent earthing points; height of the shield wire: 9 m; perfectly conducting earth.
The relative position between the lightning strike point and the distribution line in general does not influence
significantly the induced voltages if the earthing spacing is shorter than about 200 m. The situation changes, however,
in the case of larger distances. Figure 9 presents the phase-to-earth induced voltages corresponding to an earthing
spacing of 750 m and to the ideal case of earth resistance of 0 Ω. Two stroke locations are considered: in front of an
earthing point or equidistant from two adjacent earthing points. The former situation is the most favourable in terms of
the attenuation of the voltages induced on the phase conductors.
Figure 10 shows induced voltages on a 3 km long line with horizontal configuration and matched at both ends. The
lightning channel is 50 m from the line and equidistant from its terminations. The simulations refer to a subsequent
stroke current (I) of 24 kA whose waveform is obtained by the sum of two Heidler functions [64] with the following
parameters: I01 = 21.4 kA, τ11 = 0.25 µs, τ2 = 2.5 µs, I02 = 13 kA, τ12 = 2.1 µs, τ22 = 230 µs, and n1 = n2 = 2. The current
propagation velocity is assumed to be 60 % of that of light in free space. The lightning channel is vertical, 4 km long,
has no branches, and is modelled according to the Transmission Line (TL) model [58]. The height of the phase
conductors is 10 m, earth is a perfectly conducting plane, and the voltages are calculated at the point closest to the
stroke location.
In Figure 10a the phase-to-earth voltage corresponding to a neutral conductor placed at the height of 7 m and earthed
every 400 m (earth resistance of 20 Ω and lead inductance of 7 µH) is compared with the voltage that would be induced
at the same point in the absence of the neutral. As the peak value is reached well before the effect of the neutral
earthings is felt at the observation point, only the voltage waveform is changed in this case.
It is interesting to note that in some circumstances a decrease in the earthing spacing (xg) may lead to an increase in
the phase-to-neutral voltages. This is illustrated in Figure 10b, which depicts the phase-to-neutral voltages for
xg = 200 m and xg =400 m. In both cases the stroke location is equidistant from the neutral earthing points. Before the
arrival of the voltages reflected at the neutral earthings, the ratio of the phase-to-earth to the neutral-to-earth voltages is
approximately equal to the ratio between their respective heights. However, when the reflections arrive at the
observation point, the neutral-to-earth voltage decreases abruptly, causing a sudden increase in the phase-to-neutral
voltage. The greater the value of the phase-to-earth voltage at the instant in which the reflections arrive at the
observation point is, the greater the phase-to-neutral voltages will be. As the reflections corresponding to the case of
8

xg = 200 m arrive earlier than those corresponding to xg = 400 m, the phase-to-earth voltage is higher in the former
case. Additionally, the reflections are also stronger, as in this case the earthing points are closer to the stroke location
than for xg = 400 m. The net result will be a larger phase-to-neutral voltage in the case of xg = 200 m.
Fig. 9 – Induced voltages (phase-to-ground) - influence of the relative position of the stroke channel with respect to the earthing
points (adapted from [5]). Lightning current: amplitude 24.9 kA, equivalent front time 3.5 µs;
stroke location at a distance of 70 m from the distribution line; xg = 750 m; Rg = 0 Ω;
height of the shield wire: 9 m; perfectly conducting earth.
(a) (b)
Fig. 10 – Induced voltages by a subsequent stroke of 24 kA at the point closest to the lightning channel.
Stroke location is 50 m from the line; Rg = 20 Ω.
a) phase-to-earth voltages. Curve 1: neutral not present; curve 2: neutral earthed every 400 m (xg = 400 m);
b) phase-to-neutral voltages. Curve 3: xg = 200 m; curve 4: xg = 400 m
2.B.2. Surge Arresters
In order to protect effectively an unshielded MV line against direct strokes, surge arresters should be installed on all
the phases of every pole. According to the analysis performed in [1], in the case of two spans between arresters even a
line with CFO of 350 kV and earth resistance of 10 Ω would experience flashovers in about 70 % of the direct strikes.
9

Such estimate refers to a span length of 75 m and assumes that the neutral is earthed at every pole. For three spans
between arresters, the corresponding number is 80 %. However, as pointed out by McDermott et al. [13], arresters
applied to protect an unshielded line against direct strokes may have a significant failure rate due to excess energy
dissipation.
On the other hand, most of the studies agree that the application of surge arresters in MV lines can be effective in
reducing the number of flashovers caused by indirect strokes, provided that the arrester spacing is not too large. Some
possible measures for improving the reliability of a 83 km long, 13.8 kV distribution line located in a region of ground
flash density in the range of 4 to 8 flashes/(km2
.year) were investigated by Silva et al. [2]. All transformers had surge
arresters at the primary side and the neutral was earthed at poles where surge arresters were installed. Distances
between arresters varied widely depending on the region crossed by the line, ranging from only 77 m in urban sections
to the maximum of 2200 m in some rural areas. Regarding the effectivenesses of the different alternatives considered
for improving the lightning performance of this specific line, it was shown that the installation of ZnO surge arresters
on all phases with spacing of 300 m, associated with the increase of the insulation level of the rural sections of the
feeder (through the installation of 34.5 kV type pin insulators), would reduce the number of lightning outages in
approximately 50 %, with a reduction around at least 75 % of the number of faults caused by indirect strokes.
The analysis presented by Paolone et al. [7] on the mitigation effect of surge arresters on lightning induced
overvoltages was performed with a computer code that consists in an extension of the Agrawal et al coupling model
[54]. The authors considered a single conductor, 2 km long line with different arrester spacings: 2 km (only at the line
terminals), 1 km, 500 m and 200 m. Information is given about the maximum amplitude of the induced voltage along
the line as a function of the spacing between two adjacent surge arresters for the cases of lossy (conductivity of
0.001 S/m) and perfectly conducting earth. The simulations revealed a shortcoming of the effectiveness of surge
arresters when they are separated by large distances (e.g., one surge arrester every 1 km). The computed results showed
also that an important reduction of the induced overvoltages can be achieved only with a large number of arresters,
namely one surge arrester every 200 m.
According to the computer simulations presented by McDermot et al. [13], MV lines can be effectively protected
against nearby strokes by installing arresters on each phase every 360 m, although surge arresters should be installed on
every pole and on every phase to protect an unshielded distribution line against direct strokes. This is in agreement with
the conclusions obtained by Geldenhuys and Gaunt [14], from a study performed on a 10 km long rural distribution line
of 11 kV in South Africa, that surge arresters installed on the three phases with spacing of 300 m practically eliminate
outages caused by indirect discharges. A surge arrester installation interval of 200 m (on all phases) is suggested by
Yokoyama [6, 15] to give a fairly good protection for medium voltage distribution lines against overvoltages induced
by nearby strokes.
In contrast with these findings, the major conclusion of a project described by Short and Ammon [65] was that there
was little evidence that arresters used for overhead distribution line protection would provide significant improvements
in reliability. Five feeders were extensively monitored for three years. Three of them had arresters added on all phases
with spacing of 40 m (every pole), 200 m and 400 m, while the other two were used as "control circuits" and had no
arresters. During the project the circuits with arresters did not show any quantifiable improvement over the circuits
without arresters, regardless of spacing. According to the authors, even the circuit with arresters on almost every pole,
which was expected to be virtually lightning-proof, had as many lightning-caused faults as the other circuits. This is
indeed a surprising result, but the study period was not long enough to allow a definitive conclusion. Moreover, the
incidence of lightning flashovers was rather low during the monitoring period and a line survey found that about 10 %
of the arresters were missing.
An investigation about the effectiveness of surge arresters in reducing the magnitudes of lightning induced voltages
on overhead distribution lines was carried out by Piantini and Janiszewski [9]. Experiments were performed through the
use of both a reduced scale model and a full-scale system implemented specifically for this purpose, while the
simulations were conducted using the ERM [49 - 51]. The scale model was used to simulate a typical 15 kV
distribution line (horizontal configuration, height above ground of 10 m, and separation between phase conductors of
0.75 m). The MO surge arrester protective level was about 34 kV (for an impulse current of 5 kA, 8/20 µs waveform).
For the validation of the arrester model, comparisons were performed with test results corresponding to actual SiC and
ZnO distribution arresters, all with rated voltage and current of 12 kV and 5 kA, respectively. Details about the
modelling of the various system components can be found in [55]. The results showed that surge arresters can be
effectively used for improving the lightning performance of distribution lines even if they are not applied at every pole.
However, the influence of the distance between adjacent surge arresters on the induced voltages can be significant,
particularly if the lightning strike point is nearly equidistant from two sets of surge arresters. The closer the arresters
10

are, the lower the voltage magnitudes will be. Figure 11 compares the induced voltage on an 1.4 km long straight line
without surge arresters with those corresponding to arrester spacings of 300 m and 600 m. The stroke current front time
and time to half-value are about 3.2 µs and 58 µs, respectively. All the quantities are referred to the full-scale system by
applying the corresponding scale factors.
The induced voltages are in general not much affected by the relative position between the lightning strike point and
the surge arresters in case of arrester spacings shorter than about 300 m. For larger spacings, the differences between
the voltage magnitudes corresponding to the cases of stroke location in front of a set of arresters or equidistant from
two sets of arresters tend to increase, especially if the earth resistance is low. For short arrester spacings, i.e., when the
time for propagation of the reflected waves between two adjacent arresters is much smaller than the induced voltage
rise time, this behaviour is justified because the reflected waves that cause the voltage reduction vary little with the
position of the stroke channel with respect to the nearest set of arresters.
On the other hand, for large spacings the induced voltage levels are more sensitive to the distances from the stroke
location to the arresters, and the same occurs with the components (henceforth called suppressive components) that, by
coupling, cause a reduction of the voltages induced on the phase conductors. The closer the surge arrester is from the
lightning strike point, the greatest will its contribution be in reducing the voltage, so that the most favourable situation
corresponds to stroke locations nearly in front of a set of arresters. Besides, in this case the contributions of these
arresters are felt almost instantaneously, whereas if the stroke location is equidistant from two sets of arresters the
effects of the suppressive components are delayed due to the time needed for their propagation from the arresters to the
observation point. The latter situation is illustrated in Figures 11a and 11b, where the delay of the effect of the arresters
can be clearly noticed.
(a) (b)
Fig. 11 - Voltages induced on lines with and without surge arresters (s.a.), at the point closest to the stroke location, for arrester
spacings of 300 m and 600 m. Lightning strike point 70 m from the line and equidistant from
two sets of arresters (adapted from [9]). I = 38 kA; tf = 3.2 µs; Rg = 50 Ω.
a) arrester spacing equal to 300 m b) arrester spacing equal to 600 m.
The influence of the distance between adjacent surge arresters on the induced voltages is also illustrated in Figure 12
for the case of strike point 50 m from a 3 km long line and equidistant from its terminations. The line is three-phase,
without neutral, and has a horizontal configuration. The conductors are at the height of 10 m above a perfectly
conducting earth plane and the stroke current has a triangular waveform with time to half-value of 50 µs. The voltages
are calculated at the point closest to the stroke location for distances of 400 m and 200 m between adjacent surge
arresters. For comparison purposes, the voltages that would be induced at the observation point in the absence of the
arresters are also presented. Figure 12a refers to a current of 50 kA, front time of 3 µs, and propagation velocity of
30 % c. These parameters are representative of a first stroke current. Figure 12b refers to a current of 15 kA, front time
of 0.5 µs, and propagation velocity of 60 %c, which represents a typical subsequent stroke.
11

As shown in Figure 12a, when the front time is long in comparison with the time required for the reflections at the
surge arresters to arrive at the observation point, the voltage magnitudes are significantly reduced. On the other hand, if
the voltage peak value is reached before the arrival of the reflections, only the voltage wavetail will be affected, as
illustrated in Figure 12b. Therefore, the greater the stroke current steepness, the smaller the effect of the surge arresters.
(a) (b)
Fig. 12 – Induced voltages at the point closest to the lightning channel.- influence of the distance between adjacent surge arresters.
Stroke location is 50 m from the line, Rg = 50 Ω.
Curve 1: line without surge arresters; curve 2: arresters every 400 m; curve 3: arresters every 200 m.
a) I = 50 kA, tf = 3 µs, v = 30 % c b) I = 15 kA, tf = 0.5 µs, v = 60 % c
Figure 13 presents experimental results obtained from measurements carried out on the scale model considering the
more complex network configuration indicated in Figure 14. Two values were considered for the distances se and sd
between the transformer (measuring point) and the nearest set of arresters: se = 75 m and sd = 75 m, and se = 148 m and
sd = 174 m. The line was three-phase and the neutral conductor was earthed at the middle of each lateral and at all
points in which transformers or arresters were placed. The stroke current magnitude and front time were, respectively,
34 kA and 2 µs. The value of the earth resistance was 50 Ω except at the measuring point, where Rg = 0 Ω. As
expected, a significant difference is observed between the voltage amplitudes relative to the two situations.
(a) (b)
Fig. 13 - Measured phase-to-earth induced voltages considering different distances between surge arresters and measuring point.
Stroke current: amplitude 34 kA, front time tf = 2 µs. Test configuration shown in Figure 14.
a) se = 75 m; sd = 75 m b) se = 148 m; sd = 174 m
12

Fig. 14 – Test configuration related to the voltage waveforms presented in Figure 13. Neutral conductor grounded at the middle of
each lateral and at all installation points of transformers and surge arresters; r.s.m.: stroke location.
Figure 15a illustrates the influence of the return stroke current magnitude on the ratio of the peak values of the
induced voltages measured simultaneously on two distribution lines located at the same distance (70 m) from the strike
point. One of the lines has no surge arresters, while the other has MO arresters added on the three phases every 300 m.
The stroke current front time is about 3.2 µs and the voltages are measured, on each line, at the closest point to the
stroke location. The results were obtained from the scale model [9], and the ratio between the peak values of the
voltages induced on the lines with and without arresters (Uwith/Uwithout) is presented for different values of Rg.
When the lightning strike point is located in front of a set of surge arresters, the ratio Uwith/Uwithout tends to decrease
with the return-stroke current magnitude. This behaviour is more evident for low values of earth resistance, as for in
these cases the voltage on the protected line (at the point closest to the stroke location) is very close to the voltage at the
surge arrester terminals and varies little with the return stroke current peak, while the voltage on the line without
arresters is directly proportional to the current. Figure 15a shows that, for Rg greater than 50 Ω, the influence of the
return stroke current peak is not significant, i.e., the ratio between the peak values of the voltages induced on the lines
with and without surge arresters practically does not depend on the current peak for currents with the same front time.
In the case of strokes equidistant from two sets of arresters and currents with a fixed front time, the degree of
reduction of the induced voltage (i.e., the ratio Uwith/Uwithout) in general does not depend significantly upon the return
stroke current amplitude, as shown in Figure 15b. On the other hand, for currents with the same time variation rate
(dI/dt), the front time increases with the current peak I, resulting in a larger reduction of the induced voltage amplitude,
with respect to that of the unprotected line, as I increases.
The earth resistance may have a significant influence on the induced voltage amplitude, especially when the lightning
strike point is in front of a set of arresters. This is due to the fact that, for lower values of Rg, the current that flows to
earth (through the surge arresters) increases, thus increasing the value of the voltage component that, by coupling,
reduces the voltages induced on the phase conductors. For illustration, Figure 16 compares the voltage induced on a
line without arresters with those corresponding to the case of arrester spacing of 600 m, lightning strike point in front of
a set of arresters and earth resistances of 0 Ω and 200 Ω.
13

(a) (b)
Fig. 15 - Ratio of the voltages induced on lines with and without MO surge arresters (Uwith/Uwithout) as a function of the stroke current
magnitude (adapted from [9]). Stroke location is 70 m from the lines; tf = 3.2 µs; arrester spacing = 300 m.
a) lightning strike point in front of a set of arresters b) lightning strike point equidistant from two sets of arresters
Fig. 16 - Voltages induced on lines with and without surge arresters. Lightning strike point is in front of a set of arresters, at a
distance of 70 m from the line (adapted from [9]). I = 38 kA; tf = 3.2 µs; arrester spacing = 600 m.
In [9] the lightning performances of distribution lines with and without surge arresters are compared in terms of the
average annual number of lightning induced voltages with magnitudes exceeding a given insulation level. The results
refer to lines in open ground, and indicate clearly that the installation of surge arresters may improve the distribution
line performance with respect to indirect strokes. Both for the first and subsequent strokes the time to crest was
determined, in each event, from the ratio between the current magnitude and its rate of rise TV5%, which is defined as
the current rate of rise between the instants corresponding to 30 % and 90 % of the first peak value whose probability
of being exceeded is 5 %. This approach leads to conservative results. The current propagation velocity was assumed
current dependent, although the existence of a relationship between these two parameters is not generally supported by
experimental data, as pointed out by Rakov [66]. Regarding the relative position of the line and the lightning channel,
the most critical situation, i.e., lightning strike point equidistant from two sets of surge arresters, was considered in the
simulations, and this leads also to conservative results. On the other hand, earth was assumed as a perfectly conducting
plane, an assumption that may lead to induced voltage magnitudes lower than those corresponding to situations where
the finite ground conductivity is taken into account, as demonstrated by Borghetti et al. [67].
14

3 LV NETWORKS
The increasing application of sensitive electronic equipment and the increased awareness of issues concerning power
quality have highlighted the importance of improving the lightning performance of LV networks, which are much more
susceptible to lightning-caused disturbances than MV lines because of their lower withstand capability.
3.A Lightning Overvoltages
Lightning transients on low-voltage networks can be produced by several mechanisms, which can be classified into
the following categories:
- direct strikes to the LV system;
- cloud flashes;
- indirect strikes;
- transference from the medium voltage system.
3.A.1. Direct Strokes
If a flash hits a LV line, multiple flashovers occur and a rough estimation of the overvoltage can be obtained, if the
propagation effects are disregarded, by multiplying the stroke current by the equivalent earth impedance. However,
even in the case of a low value for the effective impedance, voltages much larger then the line lightning impulse
withstand level would result, which would lead to further flashovers.
In general, low-voltage networks are not that prone to direct strikes due to their relatively short lengths and to the
shielding provided by the MV line, trees and nearby structures. However, in some rural and semi-urban areas, exposed
LV lines longer than 1000 m do exist, and in case of direct lightning hits, the resulting overvoltages can damage
unprotected connected equipment.
A direct strike to the lightning protection system or to other parts of an end-user building causes an earth potential
rise that may lead to the operation of surge protective devices or to flashovers between the structure and the line
conductors. In both situations a portion of the stroke current is injected into the power line, producing overvoltages that
propagate along the network. This portion depends mainly on the relative impedance of the line with respect to the
impedances of all the other possible current paths (local earth, metallic pipes and other services such as
telecommunications lines). Figure 17 illustrates the situation, showing a case in which 50 % of the stroke current enters
the earth termination system and 50 % is distributed evenly among the services entering the structure.
Fig. 17 - Injection of surges into the low-voltage power line due to a direct strike to an end-user installation.
The current division depends on the relative impedances of the all possible current paths.
3.A.2. Cloud Flashes
Cloud discharges, which include intracloud, cloud-to-cloud, and cloud-to-air flashes, last typically between 200 ms
and 500 ms [68] and are the most frequent type of discharges, representing about 75 % of the global lightning activity
[66, 69]. Nevertheless, the number of studies conducted about this phenomenon is relatively scarce in comparison with
that bearing on cloud-to-ground flashes, which have a much greater impact in terms of deleterious effects. The main
15

practical interest in cloud flashes lies in the protection of aircraft and space craft, although the short interval between
the associated induced voltage pulses may cause degradation, damage, and failure of electronic components of sensitive
apparatus connected to the low-voltage power supply.
The voltages induced on complex low-voltage power installations by cloud (IC) and cloud-to-ground (CG)
discharges have been studied in [70 - 73]. The investigation carried out by Galván et al. [70] refers to two small
networks isolated from the power supply. Simultaneous measurements of the incident vertical lightning electromagnetic
fields and the corresponding induced voltages across a 50 Ω resistor connected between one of the phase conductors
and earth were performed. The technique proposed by the authors uses these measurements to extract the transient
response of the power installation, irrespective of its complexity. The peak-to-peak values of the four voltages induced
by cloud flashes shown in the paper are below 2 V, and in all cases a relatively good agreement was found between
measured and calculated results.
The experimental study conducted by Silfverskiöld et al. [71] compares the amplitudes of the common-mode
voltages induced on a residential installation in the complete duration of typical CG (negative and positive) and IC
flashes. The installation was disconnected from the power distribution line. The measurements of the vertical
component of the electric field and the corresponding induced voltages on the power installation showed that the
discharge events that take place inside the cloud, preceding CG and IC flashes, give rise to bipolar pulses with very fast
rise time. The pulse trains associated with such processes may induce voltages with magnitudes of the same order of
(and even higher than) those induced by the return stroke itself. These events are therefore important and should be
taken into account in the evaluation of the interference problems caused by lightning electromagnetic pulses (LEMP).
From the analysis of the obtained results the authors estimate, from a typical lightning within a distance of a few
kilometers from the low-voltage power installation network, a few tens of induced voltage pulses exceeding 400 V
peak-to-peak.
A further investigation of the transient response of low-voltage power installations to lightning electric fields was
performed by Galván and Cooray in [72], where comparisons between measured and calculated induced voltages using
the measured lightning electric field (both inside and outside the installation) as the driving source are also presented.
The amplitudes and waveforms of the induced voltages were found to be highly dependent on the soil resistivity and on
the loads connected to the low-voltage power installation. In [73] Galván et al. apply the technique proposed in [70] to
a simple circuit and to a complex wiring system. Comparisons between measured and simulated induced voltages,
presented for both systems, are in good agreement. Discussions are provided on the advantages and limitations of the
method, which represents an useful tool for evaluating induced voltages in electrical installations with linear behaviour.
Voltages induced by cloud discharges at both open-circuited terminations of an unenergized 460 m long distribution
line, together with the corresponding electric fields, are reported by Rubinstein and Uman [74]. The line consisted of
two conductors arranged in a vertical configuration, and the peak-to-peak voltages induced at the top conductor by a
flash at an altitude greater than about 5 km were around 140 V.
Even though further investigations are necessary to better characterise the voltages induced by cloud discharges as
well as the significance of their effects on sensitive loads, protection measures against the more severe types of
lightning surges are likely to be effective against such transients.
3.A.3. Indirect Strikes
When lightning strikes the earth or an object in the vicinity of a distribution network, the voltages that arise on the
low-voltage conductors may be subdivided as follows:
- voltages induced “directly”, due to the electromagnetic coupling between the line and the stroke channel;
- voltages associated with the part of the stroke current that is intercepted by the earthing points of the neutral
conductor;
- voltages transferred from the MV line.
From experiments performed at Camp Blanding using the rocket-and-wire technique to trigger lightning, Rakov and
Uman [75] and Fernandez et al. [76] showed that when the strike point is at tens of meters from the line, an appreciable
fraction of the total current enters the system from the neutral earth connections. In three cases reported in [75], in
which the distances between the line and the strike point were 60 m, 40 m and 19 m, the observed peak values of the
currents entering the system from its earth connections were, respectively, 10 %, 5 %, and 18 % of the stroke current
peak.
The voltages “directly” induced are in general the most important on account of their severity and frequency of
occurrence. However, although the number of studies conducted in this field has been increasing consistently [16, 56,
77 - 84], there is still a stark contrast to that of MV lines, to which much more attention has been given in the past.
16

The analysis carried out by Hoidalen in [78] made use of the Agrawal et al. coupling model [54] for the calculation
of lightning induced voltages on low-voltage systems. From frequency response measurements, simple models were
proposed for the input impedances of typical distribution transformers and low-voltage power installations, and their
influences upon the lightning induced voltages on simple TN and IT systems were investigated. The line considered in
the simulations was 500 m long and the phase conductors were simulated by a single wire with characteristic
impedance of 300 Ω. In one end there was a transformer, modelled as an inductance of 10 µH, while an impedance
representing the power installation was connected at the other termination. The voltage magnitudes were found to have
a high dependence on the load, the lowest values being associated with larger installations. A comprehensive
investigation was conducted by Hoidalen in [77], where the effect of the finite earth conductivity on the induced
voltages on TN and IT systems is thoroughly discussed.
In [16, 80] the authors concluded that the induced voltages are characterised by a high frequency damped oscillation
with a period equal to twice the travel time of a span (portion of the line between two adjacent neutral earthings). The
simulations were done with the LIOV-EMTP code [7, 26, 37], which is based upon the Agrawal et al. coupling model,
and considered overhead cables with two or four twisted conductors, with neutral earthing spacing in the range of
250 m to 400 m. Due to the high transient electromagnetic coupling between the conductors, for the configuration
examined the wire-to-wire voltages were disregarded and line-to-earth voltages, assumed to be the same on the
different conductors, were presented.
The influences of various parameters on the lightning induced voltages were evaluated by Piantini and Janiszewski
[81] for the case of a 300 m long, single-phase line. In [82], a line with twisted conductors was considered. The main
difference between the bunched cable and the open wire configuration is that the former is characterised by a stronger
coupling between the wires, that are much closer. The greater the mutual surge impedance between the neutral and
phase conductors, the smaller the induced voltage magnitudes. If the conductors are twisted, this impedance varies
along the line. However, as the distance between the wires is much smaller than their heights above ground, the
variation is small and for practical cases it can be neglected. The simulations, performed using the ERM [49 - 51],
showed that the induced voltages have a great impact on the lightning performance of low-voltage distribution lines.
Measurements performed by Hoidalen [83] in Norway, where the ground flash density is mostly below
1 / (km2
.year), show that more than 1000 voltages above 500 V should be expected per year in a typical rural low-
voltage overhead line with isolated neutral. Voltages up to 5 kV were recorded, and according to the study,
overvoltages can be induced by strokes more than 20 km away from the line.
Rocket triggered lightning experiments with simultaneous measurements of induced voltages on a 210 m long
overhead LV line with twisted conductors are reported by Clement and Michaud [84]. The line was connected to a
transformer at one of the ends, and to a 60 m long underground cable, terminated by LV arresters, at the other. The
stroke location was either on a tower close to the underground cable termination or on the firing area, 50 m away from
this point. The induced voltages were measured at the low-voltage transformer terminals for a total of 12 launchings.
The stroke currents varied in amplitude from 4 kA to 50 kA, and the corresponding phase-to-earth and neutral-to-earth
voltages reached maximum values in the range of 2 kV to 12 kV.
The analysis of the characteristics of such surges is of great importance, since they have a high frequency of
occurrence and can often reach large magnitudes. The severity of the induced voltages depends on many lightning
parameters and is also substantially affected by the network configuration.
3.A.4. Transference from the MV system
Lightning overvoltages on the low-voltage network can be originated from the primary circuit either in the case of a
direct hit or a nearby stroke. In both situations the transformer plays an important role in the transference mechanism.
Direct strikes to the primary circuit produce overvoltages on the LV network due to transference from the
distribution transformer and to injection of current into the neutral conductor. The latter is a consequence of the earth
potential rise caused by the flow of current through the earth resistance following the operation of MV surge arresters
and the occurrence of flashovers across the MV insulators. The transferred voltages may vary widely depending on the
strike and observation points, stroke current magnitude and waveform, and line configuration.
The overvoltages that result on a typical low-voltage distribution network in the case of direct lightning hits on the
primary line were studied in [16, 85 - 88]. A typical distribution network was considered in [85] and the voltages at
different points of the LV line were calculated in order to study the basic characteristics of the transferred surges for the
case of direct strokes to the MV line. The simulations were performed with the Alternative Transients Program (ATP)
[89], and flashovers across the MV and LV insulators were taken into account according to the "Disruptive Index
Model" [90]. It was shown that a correct representation of the distribution transformer is essential and that the well-
17

known purely capacitive PI-circuit is generally not adequate for the evaluation of transferred surges. The simulations
showed also that voltage magnitudes of some tens of kilovolts may occur at the transformer LV terminals and at the
consumers’ entrances in the absence of surge protective devices.
The influences of the stroke current magnitude and front time, earth resistance and number of low-voltage power
installations on the voltages transferred to the LV network were discussed in [86]. The results showed that the
flashovers that take place across medium and low-voltage insulators affect significantly the magnitudes and waveforms
of the transferred voltages and thus should always be taken into account. In most of the cases the larger phase-to-
neutral voltages occur at the transformer terminals. The voltages are characterised by oscillations originated from the
various reflections throughout the secondary network and therefore are strongly affected by the spacing between
adjacent earthing points. The effect of the earth resistance of the neutral conductor on the phase-to-neutral voltages is
greater than that of the earth resistance at the consumers’ entrances. For a given line, the general trend of the phase-to-
neutral voltage magnitudes is to decrease with the number of consumers.
In general, the shorter the distance between the transformer and the lightning strike point, the higher the transferred
voltages. However, the insulator flashovers tend to diminish this effect. In [87], the low-voltage power installations
were represented as pure inductances, resistances or capacitances, and were connected to the LV network through 20 m
long service drops. The reflections that occur at both terminations of the service drops led, in the situations considered,
to a decrease of the phase-to-neutral voltages at the transformer terminals when the values of the load impedance
increased. At intermediate points and at the line terminations the voltage amplitudes may increase or decrease with the
load impedance, depending on the magnitudes of the reflected waves at the various discontinuity points. Concerning
the line configuration, the results obtained in [91] showed that the overvoltages are usually much larger in the open
wire line in comparison with the bunched cable, and that the energy dissipated in the power installations may be, in
some cases, ten times higher.
The assessment of the voltages transferred from the MV line due to lightning strikes in the vicinity of the distribution
network requires the knowledge of the transformer high frequency behaviour, as well as the voltages induced at the
primary side. This subject is discussed in [92 - 94] for the case of unloaded transformers, while the presence of the low-
voltage line is considered in [56, 95].
A suitable transformer model is essential for the evaluation of transferred surges. In the well-known purely
capacitive PI-circuit, the transformer is represented by the capacitances C1 (between primary and earth), C2 (between
secondary and earth), and C12 (between primary and secondary). This circuit, however, is not adequate for evaluating
transferred surges, as it greatly overestimates the overvoltages. This is illustrated in Figure 18, which presents the
measured and calculated voltages transferred to the secondary (in open circuit) of a typical distribution transformer
(rated power 30 kVA, 13.8 kV – 220 / 127 V, delta - earthed wye connected) when a standard 1.2/50 µs impulse
voltage of 1.7 kV is applied to the primary terminals short-circuited. The measured values corresponding to the 30 kVA
transformer are C1 = 0.138 nF, C2 = 0.423 nF, and C12 = 0.305 nF, and its equivalent PI-circuit is depicted in
Figure 18a. For comparison purposes, the voltage calculated using the model presented in [96] is also shown. The
absolute value of the ratio between the peak values of the measured and calculated - using the PI-circuit - voltages is
17.5 (40 V against 700 V). It can be readily seen that not only the magnitudes, but also the voltage waveforms differ
considerably.
In realistic situations, the voltages transferred to the secondary terminals oscillate with a frequency governed by the
transformer transient response and by the configuration of the LV network, with the distance between adjacent loads
playing a major role. Even though on the primary side voltages of higher magnitudes may be induced by first strokes,
the larger transferred voltages are usually associated with the subsequent strokes. The reason is that the voltages
induced by the subsequent strokes have steeper fronts and, thus, broader frequency spectra. The difference between the
magnitudes of the voltages at the LV transformer terminals is much less than that at the primary side.
As an example, Figure 19 presents the phase-to-neutral voltages at the transformer terminals, both with and without
surge arresters at the MV side, considering a typical subsequent stroke current of 12 kA whose waveform is obtained
by the sum of two Heidler functions [64] with the following parameters: I01 = 10.7 kA, τ11 = 0.25 µs, τ2 = 2.5 µs,
I02 = 6.5 kA, τ12 = 2.1 µs, τ22 = 230 µs, and n1 = n2 = 2. The lightning channel is vertical, 4 km long, has no branches
and is modelled according to the Transmission Line (TL) model [58]. The current propagation velocity is assumed to be
60 % of that of light in free space and the stroke location is in front of the transformer, 50 m from the LV line. The LV
line is three-phase, lossless, 1000 m long, and the heights of the phase and neutral conductors are, respectively, 6.5 m
and 6.48 m. Its configuration is depicted in Figure 20. No coupling is considered between the primary and secondary
circuits, and earth is assumed as a perfectly conducting plane.
The transformer is represented by the model developed in [96], while the input impedance of the power installations
18

seen from the line are simulated according to the circuit obtained in [78] for a typical installation in the TN system. The
distance between adjacent neutral earthings is 200 m. The low-voltage power installations are close to the line and,
thus, in Figure 20b each impedance formed by an inductance of 7 µH in series with a resistance of 20 Ω corresponds to
the equivalent impedance of the earth connections of the neutral and of the closest consumer’s installation. In this way,
each equivalent impedance takes into account the inductances of the earth lead and of the service drop, as well as the
earth resistances at that neutral earthing point and at the closest service entrance.
(a) (b)
Fig. 18 - Equivalent capacitive PI-circuit and comparison of measured and calculated phase-to-neutral voltages transferred to the
secondary of a 30 kVA distribution transformer, in the no-load condition, for a standard 1.2/50 µs impulse voltage
of 1.7 kV applied to the primary terminals short-circuited.
a) equivalent capacitive PI-circuit
b) transferred voltages (curve 1: capacitive PI-circuit; curve 2: model presented in [96]; curve 3: measured)
(a) (b)
Fig. 19 - Phase-to-neutral voltages at the transformer terminals considering a typical subsequent stroke current of 12 kA. Stroke
location in front of the transformer, at a distance of 50 m from the LV line (configuration depicted in Figure 20).
Curve 1: without MV surge arresters; curve 2: with MV surge arresters.
a) induced voltages at the MV side
b) transferred voltages to the LV side
19

(a)
(b)
Fig. 20 - Low-voltage line configuration considered in the simulations presented in Figure 19. The triangles and the rectangles
denote, respectively, the distribution transformer and the low-voltage power installations.
a) top view b) side view
3.B Lightning Protection of LV Networks
The overvoltages stemmed from lightning have often high magnitudes and are usually the major cause of failures or
damage to transformers and consumers’ electric appliances, especially in the case of lines characterised by poor pole
earthing conditions and located in regions with high lightning incidence.
A fault on a transformer is always sustained and the corresponding costs are related to the repair or replacement of
the equipment and to the service interruption. Protection measures such as the application of secondary arresters and the
reduction of the earth resistance at the transformer pole can improve the lightning performance of LV networks.
However, although there has been an increasing awareness of the effectiveness of the application of low-voltage
arresters, this practice has not been widespread among electric utilities. With a few exceptions, e.g. [97], this measure is
usually taken only to meet cases where the action – mainly for improving the earth connections - did not give
satisfactory results [84] or to solve recurrent problems. This is in fact an economic issue, a trade-off between the
lightning costs and the investment in the protection scheme. The failure rate of the surge protective devices has also to
be considered in the cost-benefit analysis. However, even without taking into account the costs of damages to
consumers’ equipment, their application to transformer secondaries can be justified in areas with high lightning damage
rates, as pointed out by Darveniza [98] and by Dugan et al. [99].
Gapless secondary arresters of the metal oxide varistor (MOV) type are the most appropriate to protect the LV
network. The impact, on the lightning overvoltages, of the application of such devices to the transformer and service
entrances will be discussed in this Section.
3.B.1. Distribution Transformers
Although several factors can cause distribution transformers to fail, in lightning-prone regions failure rates can be
more than twice the norm, which is typically between 0.8 % to 1.5 % for non-interlaced and 0.4 % to 0.7 % for
interlaced transformers [100]. Most of the additional failures may be due to current surges in the low-voltage windings
[101]. These surges can be created whenever a significant portion of the stroke current is injected into the neutral
20

between the transformer and the power installations. The problems related to the so called “low-voltage side surges” or
“secondary side surges” have been discussed in [99 – 104].
Let us consider the case of a direct strike to the primary line, as illustrated in Figure 21. The MV arrester discharge
current splits so that one portion flows through the pole earth lead and another is injected into the neutral conductor.
The division of the bulk of the stroke current is determined by the earth resistances, but in the beginning of the transient
it is highly dependent on the ratio between the inductances of the two paths. Therefore, as the path to the pole earth is
shorter and the corresponding inductance is lower, initially a greater portion of the current flows through the
transformer earth lead. After a few microseconds, when the current time derivative becomes smaller, the influence of
the inductances decreases and the division is controlled by the resistances. The lower the earth resistance of the service
entrance in comparison with that of the transformer pole, the larger the magnitude of the surge current that enters the
neutral and the worse the problem.
Fig. 21 - Injection of current into the neutral due to a direct strike to the MV line.
The voltage drop across the neutral, produced by the surge current, gives rise to equal currents in the two phase
conductors - if the configuration is symmetric -, which can damage loads and cause meter gaps to flash over. These
currents flow through the transformer secondary, as indicated in Figure 21, and induce a surge voltage in the primary
which can cause part of the winding to short out. This would change the transformer ratio and subject the load to
sustained overvoltages, resulting in damage to consumers’ equipment. The surge voltage can also cause a layer-to-layer
insulation breakdown and a subsequent transformer failure and power outage. As pointed out by Dugan et al. [99], the
surge voltage distribution inside the transformer is such that the primary arrester has little effect on the prevention of
this failure.
Although for the symmetric configuration the surge current divides equally into the secondary windings, there is a
significant, equal and opposite voltage induced in each half of the MV winding so that the net voltage across the
primary terminals is nearly zero. The magnitude of this induced voltage is approximately proportional to the voltage
across the secondary windings, and it can be estimated reasonably well by considering only the inductances [99].
Hence, transformers with lower secondary impedances have a better performance against this type of surge. This is the
case when the secondary windings are interlaced, as in this condition the impedance at surge frequencies is about a
tenth that of non-interlaced transformers [100].
On the other hand, interlacing the secondary windings is not effective to solve the problem in the case of unbalanced
surges, i.e., when the currents through the secondary windings are not equal. This situation happens, for instance, if
flashover or arrester operation occurs on only one side of a service. According to Marz and Mendis [100], up to half of
all interlaced transformer failures may be due to secondary side surges. A better solution involves the application of
low-voltage arresters to the transformers, as in this case protection is provided against both balanced and unbalanced
surges, regardless of winding connection.
The situation illustrated in Figure 21 refers to just one power installation connected to the LV line. In the case of
multiple services from the transformer, a lower voltage drop will develop across the neutral and therefore less current
will be forced into the secondary transformer terminals. The stresses on both the transformer and the consumers’ loads
will then be reduced in comparison with the single service case. On the other hand, multiple services mean higher line
21

lightning exposure and a possible increase in the number of surges may counteract this effect.
Longer lengths of the LV circuit lead to greater voltage drops and consequently to current surges of higher
magnitudes impressed on the transformer secondaries. This voltage increase is however not linear, as the ratio of
resistance to inductance of the entire circuit generally increases, changing the dynamic response of the circuit and
reducing the rate of rise of the surge currents [101].
Comparisons between surge currents in secondary windings for three types of cable, namely the open wire, triplex,
and shielded, are presented in [104]. The best results, i.e., the lower currents, are obtained with the shielded cable. Due
to the greater spacing between conductors and lower mutual coupling between wires, an open wire line has higher
inductance than a triplex cable of the same length, and therefore a larger net voltage develops across the neutral,
causing a surge current of higher magnitude to flow in the transformer and consumers’ loads.
Long-term studies performed by Darveniza and Mercer [97] led to an improved lightning protection scheme for
exposed pole mounted transformers in Australia. The protection measures consisted in the relocation of primary surge
arresters close to the terminals and in the fitting of secondary arresters. The lightning damage rates decreased from
about 2 to 0.3 per 100 transformers per year after the implementation of the protection system [98], and the authors
attribute this reduction mainly to the installation of secondary arresters.
It is important to note that protecting the transformer by means of interlaced secondaries or low-voltage arresters
results in larger surge currents, as both measures provide a low impedance for the surge. As a consequence, customers’
devices may be subjected to higher voltage stresses.
Concerning the transformer LV arrester, discharge levels of 2 kV to 5 kV are considered adequate. Although the
lightning impulse withstand level of a transformer secondary is typically in the range of 20 kV to 30 kV, insulation
degradation may be caused by overloading, so that lower protective levels may be beneficial. Secondary arrester classes
between 175 V and 650 V are in principle suitable, but 440 V or 480 V arresters have some advantages over both the
limits. They have better coordination with the primary arrester than 175 V arresters, which are susceptible to thermal
failures caused by switching events [100, 101, 103], and have a lower discharge voltage than a 650 V arrester, thus
reducing the risk of damage to sensitive consumers’ devices.
As pointed up by Dugan et al. [99], the LV transformer arresters must not substitute one type of failure for another
and should be designed so that the possibility of failure is remote. Although the magnitude of a typical secondary
current surge may be less than 1 kA, the arresters should have a current discharge capability of at least half of that for a
standard distribution class arrester, i.e., in the range of 20 kA to 40 kA.
In the simulations performed in the next section, the transformer is protected with 440 V secondary gapless MOV
arresters, whose equivalent circuit is a capacitance of 780 pF in parallel with a non-linear resistor with the V/I
characteristic depicted in Figure 22. The inductance of the connecting leads is disregarded.
Fig. 22 - V/I characteristic of the secondary MOV arresters.
3.B.2. Low-voltage Power Installations
The characteristics of the lightning overvoltages on the secondary network depend on a number of parameters, but in
general those induced by nearby strokes or transferred from the medium voltage line due to direct strikes to the primary
22

conductors play a more significant role in the network performance. These overvoltages may have a damaging effect on
the customers’ loads, and the use of properly coordinated surge protective devices at the service entrance and at
susceptible equipment is recommended especially in regions with high lightning incidence.
The arresters applied at the service entrance should be similar in rating to the transformer arrester and have a
discharge voltage of less than 2 kV. This level, which is already too high for sensitive electronic equipment, may be
much higher within the installation due to voltage oscillations caused by reflections at various points. Therefore, local
protection is required for such loads.
The arresters at the service entrance should have higher energy handling capability than those placed at internal parts
of the premises - sometimes referred to as “suppressors” - and divert the bulk of the surge current. None of the
protective devices should be overloaded, and this is the concept of arrester coordination. Besides the energy absorption
capability, the clamping voltages and the distances between the secondary arresters and the suppressors should also be
considered for achieving a successful coordination. This topic is specifically addressed in [105 – 109], and guidelines
for installing surge protective devices are given in [110].
In order to evaluate how the overvoltages induced at the service entrances by indirect strokes are affected by the
application of secondary arresters, the following line configurations, indicated in Figure 23, will be considered:
1) arresters not installed (configuration 1);
2) arresters only at the transformer terminals (configuration 2);
3) arresters at the transformer terminals and at all service entrances, excepting at Point 4 (configuration 3).
Fig. 23 - Low-voltage line configurations. The triangle and the rectangles denote, respectively, the distribution transformer
and the low-voltage power installations. Configuration 1: line without arresters;
configuration 2: arresters only at Point 3; configuration 3: arresters at points 1, 2, 3, 5, and 6.
The arresters at the service entrances and at the transformer are assumed to have the same characteristics. The three-
phase line is represented by an equivalent single conductor, so that the corresponding arrester capacitance to be
considered in the simulations is 2.34 nF. Likewise, the values of currents in the horizontal axis of Figure 22 should be
multiplied by 3. The transformer input impedance is simulated by a simple inductance of 16 µH, which corresponds to
the inductance of the equivalent single-phase model of the 30 kVA transformer considered in [96]. Due to the presence
of the secondary arresters at its LV terminals, the transformer model does not drastically influence the results.
The circuit adopted in Section 3.A.4 to represent the impedance seen by the power line at the service entrances is
typical of a relatively large power installation [78]. The equivalent load impedance has been shown to significantly
affect the induced voltages, larger impedances leading to higher voltages. A more conservative condition is considered
henceforth, and the loads, in the TN system, are simulated by just an inductance of 10 µH, which is representative of
smaller installations [77]. The other parameters remain the same as in Section 3.A.4.
The most critical situation, from the point of view of a power installation, is that in which the stroke location is in
front of it. This case is illustrated in Figure 24, in which the phase-to-neutral induced voltages at Point 4, relative to the
line configurations 1 and 2, are compared. The results refer to the subsequent stroke, which induces considerably higher
voltages than the first stroke for typical secondary line configurations. If the stroke location is in front of an unprotected
service entrance, the presence of arresters at other points of the line will not be effective in reducing the induced
voltage magnitude at that point. For the situation considered, the induced voltage peak value is about twice the
recommended protective level of 2 kV.
23

A more favourable situation occurs when the lightning strike point is in front of the transformer, as indicated in
Figure 23. The phase-to-neutral voltages induced by the subsequent stroke at points 2, 3, 4, and 5 are presented in
Figure 25 for the line configuration 2, which corresponds to arresters installed only at the transformer terminals. In this
case all the voltage magnitudes are lower than 2 kV.
Fig. 24 - Phase-to-neutral induced voltages at Point 4 of Figure 23 when the stroke location is in front of it, at a distance of 50 m.
I = 12 kA. Curve 1: configuration 1; curve 2: configuration 2.
(a) (b)
Fig. 25 - Phase-to-neutral induced voltages at points 2, 3, 4, and 5 of Figure 23 considering a typical subsequent stroke current
(I = 12 kA). Configuration 2 (arresters only at the transformer terminals).
a) Points 2 (curve 2) and 4 (curve 4) b) Points 3 (curve 3) and 5 (curve 5)
Let us now consider a more severe condition, in which the subsequent stroke current has the same waveform, but its
magnitude is 29.2 kA. This value has a probability of only 5 % of being exceeded [59], and the corresponding
maximum time derivative is about 120 kA/µs. Figure 26 shows the voltages induced at points 2, 3, 4, and 5. In this
case, the level of 2 kV is exceeded at all the service entrances.
The application of secondary arresters to a power installation can effectively reduce the local overvoltages to
acceptable limits. However, in some circumstances this may result in higher voltage stresses at unprotected premises.
This situation is illustrated in Figure 27, which depicts the phase-to-neutral induced voltages corresponding to the line
configuration 3. At all service entrances, with the exception of the unprotected one (Point 4), the voltages are kept
below approximately 1.1 kV. A comparison of Figures 27a and 26a shows that the voltage magnitude at Point 4 is
24

indeed larger than that relative to configuration 2 (5 kV against 3.5 kV), in which the arresters were placed just at the
transformer terminals. For the network configuration 3 and the “severe” subsequent stroke current considered, only
when the distance between the line and the stroke location is greater than 200 m the peak voltage at Point 4 will be
lower than 2 kV. Thus, in order to protect low-voltage power installations against lightning overvoltages, properly rated
and coordinated surge protective devices should be installed at all service entrances.
In comparison with TN systems, IT systems are in general subject to much larger induced voltages and are also far
more affected by the finite earth conductivity. A meticulous analysis of this topic is presented by Hoidalen in [77].
Higher values of the soil resistivity usually lead to larger voltage magnitudes, particularly in the case of IT systems. On
the other hand, as the voltage front time, which has a remarkable influence on the effectiveness of the surge protective
devices, also increases, the net result may be, in some cases, a decrease of the voltage magnitude.
(a) (b)
Fig. 26 - Phase-to-neutral induced voltages at points 2, 3, 4, and 5 of Figure 23 considering a subsequent stroke current with
amplitude I = 29.2 kA. Configuration 2 (arresters only at the transformer terminals).
(a) (b)
Fig. 27 - Phase-to-neutral induced voltages at points 2, 3, 4, and 5 of Figure 23 considering a subsequent stroke current with
amplitude I = 29.2 kA. Configuration 3 (arresters at points 1, 2, 3, 5, and 6).
When lightning strikes the MV network, short duration pulses of several tens of kilovolts may be transferred to the
secondary circuit either by the first and subsequent strokes. As in the case of nearby lightning, higher voltage
amplitudes are usually related to the subsequent stroke. The presence of arresters at various places of the LV line does
25

not prevent high voltages from arising at unprotected points.
4 CONCLUSIONS
The magnitudes and waveforms of the lightning overvoltages depend considerably on many line and stroke
parameters, which may combine in an infinite variety of ways.
The basic measures that can be applied to improve the lightning performance of MV distribution lines involve the
increase of the line insulation withstand capability, the use of periodically grounded shield wires, and the installation of
surge arresters along the line.
The frequency of flashovers associated with indirect strokes decrease substantially as the critical flashover
overvoltage of the line structures increases, and can be virtually eliminated if the CFO is greater than 300 kV.
However, the number of line faults caused by direct strokes will remain practically the same and surges with higher
magnitudes will travel over long distances, increasing the stresses on line equipment.
A shield wire may reduce the amplitudes of the overvoltages associated with nearby strokes regardless of its position
with respect to the phase conductors. However, the closer it is to the phase conductors, the greater the voltage reduction
will be. Its effectiveness decreases with the increase of both the earthing spacing and the earth resistance.
The application of line arresters can be effective in reducing the number of flashovers caused by indirect strokes,
provided that the arrester spacing is not too large.
In order to mitigate the effects of direct strokes, the shield wire should not only be earthed at every pole, but also the
line should have sufficient CFO between the earth lead and the phase conductors, and the earth resistances should be
low. Alternatively, surge arresters should be installed on all phases at very short intervals, but if the line is unshielded
the arresters may have a significant failure rate due to excess energy dissipation. Undoubtedly the best solution consists
in the use of a shield wire in conjunction with arresters on every pole and every phase. This solution is effective against
direct strokes and theoretically eliminates flashovers, but a cost-benefit analysis should always be performed.
Secondary systems are much more susceptible to lightning-caused disturbances than MV lines because of their lower
withstand capability. Larger overvoltages are usually associated with subsequent strokes, although severe surges can
also be produced by the first stroke.
Phase-to-ground voltages induced by nearby lightning can reach some tens of kilovolts in various points along the
network, especially if the stroke location is not in front of a neutral earthing point. Lower magnitudes are observed at
the transformer and customers’ entrances, but the value of 10 kV may often be exceeded in the case of strikes closer
than about 50 m. Phase-to-neutral voltages of some kilovolts are common if surge protective devices are not applied. In
the case of direct strikes to the MV line, short duration pulses of several tens of kilovolts are transferred to the
secondary circuit.
In regions of high lightning activity, surges originated in the low-voltage side can be responsible for a great number
of transformer failures or damages, even if arresters are placed close to the primary terminals. The application of
arresters on transformer secondaries can significantly reduce the lightning damage rates of exposed transformers, but it
does not prevent overvoltages from arising at the service entrances.
Similarly, the application of secondary arresters to a power installation can effectively reduce the local overvoltages
to acceptable limits, but in some circumstances this may result in higher voltage stresses at unprotected premises.
Therefore, unless they are applied at every service entrance, exposed sensitive electronic equipment can be damaged. In
fact, voltage oscillations caused by reflections at various points within the installation can give rise to internal
overvoltages with higher magnitudes than that limited by the arresters placed at the service entrance. Therefore, local
protection is required for such susceptible loads.
5 REFERENCES
[1] IEEE Std 1410, “IEEE Guide for improving the lightning performance of electric power overhead distribution lines”, July 2004.
[2] S. P. da Silva, A. Piantini, J. L. De Franco and J. Gonçalves, "Lightning performance studies for a 13.8 kV distribution network", Proceedings
of the VII International Symposium on Lightning Protection, São Paulo, pp. 137-143, Nov. 2003.
[3] Y. Asaoka, A. Asakawa, S. Yokoyama, K. Nakada and H. Sugimoto, "Study of lightning protection effects for distribution lines with lightning
protective devices attached to pole transformers", Proceedings of the VII International Symposium on Lightning Protection, São Paulo, pp. 119-
125, Nov. 2003.
[4] A. Borghetti, C. A. Nucci, M. Paolone and M. Bernardi, "Effect of the lateral distance expression and of the presence of shielding wires on the
evaluation of the number of lightning induced voltages", Proceedings of the 25th International Conference on Lightning Protection, Rhodes,
vol. A, pp. 340-345, Sep. 2000.
[5] A. Piantini and J. M. Janiszewski, "Lightning induced voltages on overhead lines, the effect of ground wires", Proceedings of the 22nd
International Conference on Lightning Protection, Budapest, pp. R3b/1-R3b/5, Sep. 1994.
[6] S. Yokoyama, "Designing concept on lightning protection of overhead power distribution line", Proceedings of the IX International Symposium
on Lightning Protection, Foz do Iguaçu, pp. 647-662, Nov. 2007.
26

[7] M. Paolone, C. A. Nucci, E. Petrache and F. Rachidi, "Mitigation of lightning-induced overvoltages in medium voltage distribution lines by
means of periodical grounding of shielding wires and of surge arresters, modeling and experimental validation", IEEE Transactions on Power
Delivery, vol. 19, no. 1, pp. 423-431, 2004.
[8] S. Yokoyama and A. Asakawa, “Experimental Study of Response of Power Distribution Lines to Direct Lightning Hits", IEEE Transactions on
Power Delivery, vol. 4, no. 4, pp. 2242-2248, Oct., 1989.
[9] A. Piantini and J. M. Janiszewski, “The effectiveness of surge arresters on the mitigation of lightning induced voltages on distribution lines”,
Journal of Lightning Research, vol. 2, pp. 34-52, 2007.
[10] A. Piantini and J. M. Janiszewski, "Protection of distribution lines against indirect lightning strokes through the use of surge arresters", CIGRE
Surge Arrester Tutorial, Rio de Janeiro, Apr. 2005.
[11] A. Piantini, T. O. de Carvalho, A. Silva Neto, J. M. Janiszewski, R. A. C. Altafim and A. L. T. Nogueira, "A System for simultaneous
measurements of lightning induced voltages on lines with and without arresters", Proceedings of the 27th International Conference on
Lightning Protection, Avignon, vol. 1, pp. 297-302, Sep. 2004.
[12] A. Piantini and J. M. Janiszewski, "Use of surge arresters for protection of overhead lines against nearby lightning", Proceedings of the 10th
International Symposium on High Voltage Engineering, Montréal, vol. 5, pp. 213-216, Aug. 1997.
[13] T. E. McDermot, T. A. Short and J. G. Anderson, "Lightning protection of distribution lines", IEEE Transactions on Power Delivery, vol. 9, no.
1, pp. 138-152, Jan. 1994.
[14] H. J. Geldenhuys and C. T. Gaunt, "Insulation co-ordination of rural distribution lines", Proceedings of the II Seminário Internacional de
Distribuição de Energia Elétrica, Belo Horizonte, 1990.
[15] S. Yokoyama, "Distribution surge arrester behaviour due to lightning induced voltages", IEEE Transactions on Power Delivery, vol. 1, no. 1,
pp. 171-178, Jan. 1986.
[16] C. Mirra, A. Porrino, A. Ardito and C. A. Nucci, “Lightning overvoltages in low-voltage networks”, Proceedings of the International
Conference on Electricity Distribution (CIRED), Birmingham, Conference Publication no. 438, pp. 2.19.1-2.19.6, June 1997.
[17] CIGRE-CIRED JWG C4.4.02, “Protection of MV and LV Networks against Lightning. Part I, Common Topics", CIGRE Technical Brochure
no. 287, 2005.
[18] C. A. Nucci and F. Rachidi, Interaction of electromagnetic fields generated by lightning with overhead electrical networks, The Lightning
Flash, ed. V. Cooray, pp. 425-478, IEE Power and Energy Series 34, The Institution of Electrical Engineers, London, 2003.
[19] V. A. Rakov and M. A. Uman, Deleterious effects of lightning and protective techniques, Lightning - Physics and Effects, pp. 588-641,
Cambridge University Press, Cambridge, 2003.
[20] C. A. Nucci et al., “Lightning-induced voltages on overhead power lines. Part III, Sensitivity analysis”, Electra, pp. 27-30, Oct. 2005.
[21] A. Borghetti, T. Henriksen, T. M. Rojas, C. A. Nucci, M. Paolone, F. Rachidi and H. Soibelzón, “Lightning protection of medium voltage
networks”, Proceedings of the 8th International Symposium on Lightning Protection (SIPDA), São Paulo, pp. 17-23, Nov. 2005.
[22] A. Piantini and J. M. Janiszewski, “The effectiveness of surge arresters on the mitigation of lightning induced voltages on distribution lines”,
Proceedings of the 8th International Symposium on Lightning Protection (SIPDA), São Paulo, pp. 777-798, Nov. 2005.
[23] V. Shostak, W. Janischewskyj, F. Rachidi, A. M. Hussein, J. L. Bermudez, J. S. Chang and S. Sokolovsky, “Estimation of lightning-caused
stresses in a MV distribution line”, Proceedings of the 27th International Conference on Lightning Protection (ICLP), Avignon, vol. 2, pp. 678-
683, Sep. 2004.
[24] S. Yokoyama, “Lightning protection of MV overhead distribution lines”, Proceedings of the 7th International Symposium on Lightning
Protection (SIPDA), São Paulo, pp. 485-507, Nov. 2003.
[25] A. Piantini, and J. M. Janiszewski, “Lightning induced voltages on distribution transformers, the effects of line laterals and nearby buildings”,
Proceedings of the 6th International Symposium on Lightning Protection (SIPDA), Santos, pp. 77-82, Nov. 2001.
[26] C. A. Nucci, A. Borghetti, A. Piantini, and J. M. Janiszewski, “Lightning-induced voltages on distribution overhead lines, comparison between
experimental results from a reduced-scale model and most recent approaches”, Proceedings of the 24th International Conference on Lightning
Protection (ICLP), Birmingham, vol. 1, pp. 314-320, Sep. 1998.
[27] M. Darveniza, “A practical approach to lightning protection of sub-transmission and distribution systems”, Proceedings of the 5th International
Symposium on Lightning Protection (SIPDA), São Paulo, pp. 365-372, May 1999.
[28] Joint CIRED/CIGRÉ Working Group 05, “Protection of MV and LV networks against lightning, Part II, application to MV networks”,
Proceedings of the International Conference on Electricity Distribution (CIRED), Birmingham, Conference Publication No. 438, pp. 2.28.1-
2.28.8, 1997.
[29] P. P. Baker, T. A. Short, A. R. Eybert-Berard and J. P. Berlandis, “Induced voltage measurements on an experimental distribution line during
nearby rocket triggered lightning flashes”, IEEE Transactions on Power Delivery, vol. 11, no. 2, pp. 980-995, 1996.
[30] A. Piantini and J. M. Janiszewski, “An Experimental study of lightning induced voltages by means of a scale model”, Proceedings of the 21st
International Conference on Lightning Protection (ICLP), Berlin, pp. 195-199, Sep. 1992.
[31] S. Rusck, “Induced lightning over-voltages on power-transmission lines with special reference to the over-voltage protection of low-voltage
networks”, Transactions of the Royal Institute of Technology, no. 120, 1958.
[32] C. A. Nucci, “Lightning-induced voltages on distribution systems, influence of ground resistivity and system topology”, Proceedings of the 8th
International Symposium on Lightning Protection (SIPDA), São Paulo, pp. 761-773, Nov. 2005.
[33] Y. Baba and V. A. Rakov, “On calculating lightning-induced overvoltages in the presence of a tall strike object”, Proceedings of the 8th
[34] A. Piantini, T. O. de Carvalho, A. Silva Neto, J. M. Janiszewski, R. A. C. Altafim and A. L. T. Nogueira, “A system for lightning induced
voltages data acquisition - preliminary results”, Proceedings of the 7th International Symposium on Lightning Protection (SIPDA), São Paulo,
pp. 156-161, Nov. 2003.
[35] C. A. Nucci, S. Guerrieri, M. T. Correia de Barros and F. Rachidi, “Influence of corona on the voltages induced by nearby lightning on
overhead distribution lines”, IEEE Transactions on Power Delivery, vol. 15, no. 4, pp. 1265-1273, 2000.
[36] A. Borghetti, C. A. Nucci, M. Paolone and F. Rachidi, “Characterization of the response of an overhead line to lightning electromagnetic
fields”, Proceedings of the 25th International Conference on Lightning Protection (ICLP), Rhodes, vol. A, pp. 223-228, Sep. 2000.
[37] C. A. Nucci and F. Rachidi, “Lightning induced overvoltages”, IEEE Transmission and Distribution Conference, Panel Session “Distribution
Line Protection”, New Orleans, Apr. 1999.
[38] A. Piantini and J. M. Janiszewski, “The influence of the upward leader on lightning induced voltages”, Proceedings of the 23rd International
Conference on Lightning Protection (ICLP), Florence, vol. 1, pp. 352-357, Sep. 1996.
[39] C. A. Nucci, F. Rachidi, M. Ianoz and C. Mazzetti, “Comparison of two coupling models for lightning-induced overvoltage calculations”, IEEE
Transactions on Power Delivery, vol. 10, no. 1, pp. 330-339, 1995.
27

[40] M. Ishii, K. Michishita, Y. Hongo and S. Ogume, “Lightning-induced voltage on an overhead wire dependent on ground conductivity”, IEEE
[41] C. A. Nucci, F. Rachidi, M. Ianoz and C. Mazzetti, “Lightning-induced overvoltages on overhead lines”, IEEE Transactions on
Electromagnetic Compatibility, vol. 35, no. 1, pp. 75-86, 1993.
[42] P. Chowdhuri, “Response of overhead lines of finite length to nearby lightning strokes”, IEEE Transactions on Power Delivery, vol. 6, no. 1,
pp. 343-351, 1991.
[43] P. Chowdhuri, “Estimation of flashover rates of overhead power distribution lines by lightning strokes to nearby ground”, IEEE Transactions
on Power Delivery, vol. 4, no. 3, pp. 1982-1989, 1989.
[44] A. C. Liew and S. C. Mar, “Extension of the Chowdhuri - Gross model for lightning induced voltage on overhead lines”, IEEE Transactions on
Power Systems, vol. 1, no. 2, pp. 240-247, 1986.
[45] S. Yokoyama, K. Miyake, H. Mitani and A. Takanishi, “Simultaneous measurement of lightning induced voltages with associated stroke
currents”, IEEE Transactions on Power Apparatus and Systems, vol. 102, no. 8, pp. 2420-2427, 1983.
[46] S. Yokoyama, K. Miyake, H. Mitani and N. Yamazaki, “Advanced observations of lightning induced voltage on power distribution lines”,
IEEE Transactions on Power Delivery, vol. 1, no. 2, pp. 129-139, 1986.
[47] S. Yokoyama, K. Miyake and S. Fukui, “Advanced observations of lightning induced voltage on power distribution lines (II)”, IEEE
[48] A. Piantini and J. M. Janiszewski, “Analysis of three different theories for computation of induced voltages on distribution lines due to nearby
lightning”, Proceedings of the International Conference on Electricity Distribution (CIRED), Buenos Aires, pp. Session 1 / 127-132, Dec. 1996.
[49] A. Piantini and J. M. Janiszewski, “Induced voltages on distribution lines due to lightning discharges on nearby metallic structures”, IEEE
Transactions on Magnetics, vol. 34, no. 5, pp. 2799-2802, 1998.
[50] A. Piantini and J. M. Janiszewski, “The Extended Rusck Model for calculating lightning induced voltages on overhead lines”, Proceedings of
the 7th International Symposium on Lightning Protection (SIPDA), Curitiba, pp. 151-155, Nov. 2003.
[51] A. Piantini and J. M. Janiszewski, “An improved model for lightning induced voltages calculations”, Proceedings of the IEEE/PES
Transmission & Distribution Conference and Exposition, Latin America, São Paulo, pp. 554-559, Nov. 2004.
[52] V. Cooray, “Calculating lightning-induced overvoltages in power lines, a comparison of two coupling models”, IEEE Transactions on
[53] K. Michishita and M. Ishii, “Theoretical comparison of Agrawal´s and Rusck´s field-to-line coupling models for calculation of lightning-
induced voltage on an overhead wire”, IEE Transactions of Japan, vol. 117, no. 9, pp. 1315-1316, 1997.
[54] A. K. Agrawal, H. J. Price and S. Gurbaxani, “Transient response of a multiconductor transmission line excited by a nonuniform
electromagnetic field”, IEEE Transactions on Electromagnetic Compatibility, vol. EI-22, pp. 119-129, May 1980.
[55] A. Piantini, J. M. Janiszewski, A. Borghetti, C. A. Nucci and M. Paolone, “A scale model for the study of the LEMP response of complex
power distribution networks”, IEEE Transactions on Power Delivery, vol. 22, no. 1, pp. 710-720, 2007.
[56] A. Piantini, “Lightning transients on LV networks caused by indirect strokes”, Journal of Lightning Research, vol. 1, pp. 111-131, 2007.
[57] Y. Baba and V. A. Rakov, “Voltages induced on an overhead wire by lightning strikes to a nearby tall grounded object”, IEEE Transactions on
[58] M. A. Uman and D. K. McLain, “Magnetic field of the lightning return stroke”, Journal of Geophysical Research, vol. 74, pp. 6899-6910,
1969.
[59] R. B. Anderson and A. J. Eriksson, “Lightning parameters for engineering application”, Electra, no. 69, pp. 65-102, 1979.
[60] S. Yokoyama, "Experimental analysis of earth wires for induced lightning surges", IEE Proceedings -C, vol. 127, no. 1, pp. 33-40, 1980.
[61] S. Yokoyama et al., "Analogue simulation of lightning induced voltages and its application for analysis of overhead-ground-wire effects", IEE
Proceedings -C, vol. 132, no. 4, pp. 208-216, 1985.
[62] P. Chowdhuri and E. T. B. Gross, "Voltages induced on overhead multiconductor lines by lightning strokes", Proceedings of the IEE, Vol. 116,
no. 4, pp. 561-565, 1969.
[63] E. Cinieri and A. Fumi, "Effetto de11a presenza di piú conduttori e di funi di guardia su11e sovratensioni atmosferiche indotte ne11e 1inee
e1ettriche", L” Energie E1ettrica, no. 11/12, pp. 595-601, 1979.
[64] F. Heidler, “Analytische Blitzstromfunktion zur LEMP- Berechnung”, Proceedings of the 18th International Conference on Lightning
Protection (ICLP), Munich, , pp. 63-66, Sep. 1985.
[65] T. A. Short and R. H. Ammon, "Monitoring results of the effectiveness of surge arrester spacings on distribution line protection", IEEE
Transactions on Power Delivery, vol. 14, no. 3, pp. 1142-1150, July 1999.
[66] V. A. Rakov, “Lightning phenomenology and parameters important for lightning protection”, Proceedings of the 9th International Symposium
on Lightning Protection (SIPDA), Foz do Iguaçu, pp. 541-564, Nov. 2007.
[67] A. Borghetti, C. A. Nucci, M. Paolone, M. Bernardi, S. Malgarotti, I. Mastandrea “Influence of surge arresters on the statistical evaluation of
lightning performance of distribution lines”, Proceedings of the International Conference on Probabilistic Methods Applied to Power Systems,
2004.
[68] V. Cooray, The mechanism of the lightning flash, The Lightning Flash, ed. V. Cooray, pp. 127-239, IEE Power Engineering Series 34, The
Institution of Electrical Engineers, London, 2003.
[69] V. A. Rakov and M. A. Uman, Cloud discharges, Lightning - Physics and Effects, pp. 321-345, Cambridge University Press, Cambridge, 2003.
[70] A. Galván, V. Cooray, and R. Thottappillil, “Effects of lightning electromagnetic field pulses (LEMP) in low-voltage power installations”,
Proceedings of the 24th International Conference on Lightning Protection (ICLP), Birmingham, pp. 729-734, Sep. 1998.
[71] S. Silfverskiold, R. Thottappillil, V. Cooray, and V. Scuka, “Induced voltages in a low-voltage power installation network due to lightning
electromagnetic fields: an experimental study”, IEEE Transactions on Electromagnetic Compatibility, vol. 41, no. 3, pp. 265-271, 1999.
[72] A. Galván and V. Cooray: “Analytical simulation of lightning induced voltage in low-voltage power installations”, Proceedings of the 25th
International Conference on Lightning Protection (ICLP), Rhodes, pp. 290-295, Sep. 2000.
[73] A. Galván, V. Cooray, and R. Thottappillil, “A Technique for the evaluation of lightning-induced voltages in complex low-voltage power-
installation networks’, IEEE Transactions on Electromagnetic Compatibility, vol. 43, no. 3, pp. 402-409, 2001.
[74] M. Rubinstein and M. A. Uman, “Review of the University of Florida research on lightning induced voltages on power distribution lines”,
Proceedings of the 21st International Conference on Lightning Protection (ICLP), Berlin, pp. 189-193, Sep. 1992.
[75] V. A. Rakov and M. A. Uman, Artificial initiation (triggering) of lightning by ground-based activity, Lightning - Physics and Effects, pp. 265-
307, Cambridge University Press, Cambridge, 2003.
[76] M. I. Fernandez, V. A. Rakov, and M. A. Uman, “Transient currents and voltages in a power distribution system due to natural lightning”,
Proceedings of the 24th International Conference on Lightning Protection (ICLP), Birmingham, pp. 622-629, Sep. 1998.
28

[77] H. K. Hoidalen, “Lightning-induced overvoltages in low-voltage systems”, PhD Thesis, Norwegian University of Science and Technology,
1997.
[78] H. K. Hoidalen, “Lightning-induced voltages in low-voltage systems and its dependency on overhead line terminations”, Proceedings of the
24th International Conference on Lightning Protection (ICLP), Birmingham, pp. 287-292, Sep. 1998.
[79] A. De Conti and S. Visacro, “Calculation of lightning-induced voltages on low-voltage distribution networks”, Proceedings of the 7th
[80] Joint CIRED/CIGRÉ Working Group 05, “Protection of MV and LV networks against lightning. Part I: basic information”, Proceedings of the
International Conference on Electricity Distribution (CIRED), Birmingham, Conference Publication No. 438, pp. 2.21.1 - 2.21.7, June 1997.
[81] A. Piantini and J. M. Janiszewski, “Lightning induced overvoltages on low-voltage lines”, Proceedings of the 5th International Symposium on
Lightning Protection (SIPDA), São Paulo, pp. 234-239, May 1999.
[82] A. Silva Neto and A. Piantini, “Induced overvoltages on LV lines with twisted conductors due to indirect strokes”, Proceedings of the 8th
[83] H. K. Hoidalen, “Lightning-induced overvoltages in rural isolated neutral low-voltage systems”, Proceedings of the 28th International
Conference on Lightning Protection (ICLP), Kanazawa, pp. 485-490, Sep. 2006.
[84] M. Clement and J. Michaud, “Overvoltages on the low voltage distribution networks. Origins and characteristics. Consequences upon the
construction of Electricite de France networks”, Proceedings of the 12th International Conference on Electricity Distribution (CIRED),
Birmingham, pp. 2.16.1-2.16.6, 1993.
[85] A. Piantini, A. G. Kanashiro, and P. F. Obase, “Lightning surges transferred to the low-voltage network”, Proceedings of the 7th International
Symposium on Lightning Protection (SIPDA), Curitiba, pp. 216-221, Nov. 2003.
[86] P. F. Obase, A. Piantini, and A. G. Kanashiro: “Overvoltages transferred to low-voltage networks due to direct strokes on the primary”,
Proceedings of the 8th International Symposium on Lightning Protection (SIPDA), São Paulo, Nov. 2005, pp. 489-494
[87] P. F. Obase, A. Piantini, and A. G. Kanashiro: ‘Overvoltages on LV networks associated with direct strokes on the primary line’. Proceedings
of the 28th International Conference on Lightning Protection (ICLP), Kanazawa, vol. 1, pp. 479-484, Sep. 2006.
[88] P. F. Obase, A. Piantini, and A. G. Kanashiro, “Overvoltages on distribution networks caused by direct lightning hits on the primary line”,
Proceedings of the International Conference on Grounding and Earthing and International Conference on Lightning Physics and Effects, Belo
Horizonte, pp. 198-203, Nov. 2004.
[89] ATP, Alternative transients program rule book. Leuven EMTP Center, 1987.
[90] M. Darveniza and A. E. Vlastos, “The generalized integration method for predicting impulses volt-time characteristics for non-standard wave
shapes – a theoretical basis”, IEEE Transactions on Electrical Insulation, vol. 23, no. 3, pp. 373-381, 1988.
[91] A. de Conti. and S. Visacro, “Evaluation of lightning surges transferred from medium voltage to low-voltage networks”, IEE Proc.-Gener.
Transm. Distrib., vol. 152, no. 3, pp. 351-356, 2005.
[92] A. Borghetti, R. Iorio, C. A. Nucci, and P. PELACCHI, “Calculation of voltages induced by nearby lightning on overhead lines terminated on
distribution transformers”, Proceedings of the International Conference on Power Systems Transients (IPST), Lisbon, pp. 311-316, Sep. 1995.
[93] A. Piantini and C. V. S. Malagodi, “Voltage surges transferred to the secondary of distribution transformers”, Proceedings of the 11th
International Symposium on High Voltage Engineering (ISH), London, 1, pp. 1365-1368, Aug. 1999.
[94] A. Piantini and C. V. S. Malagodi, “Voltages transferred to the low-voltage side of distribution transformers due to lightning discharges close
to overhead lines”, Proceedings of the 5th International Symposium on Lightning Protection (SIPDA), São Paulo, pp. 201-205, May 1999.
[95] A. Borghetti, A. S. Morched, F. Napolitano, C. A. Nucci, and M. Paolone, “Lightning-induced overvoltages transferred from medium-voltage
to low-voltage networks”, Proceedings of the IEEE St. Petersburg Power Tech, St. Petersburg, June 2005
[96] A. Piantini and A. G. Kanashiro, “A distribution transformer model for calculating transferred voltages”, Proceedings of the 26th International
Conference on Lightning Protection (ICLP), Cracow, vol. 2, pp. 429-434, Sep. 2002.
[97] M. Darveniza and D. R. Mercer, “Lightning protection of pole mounted transformers”, IEEE Transactions on Power Delivery, vol. 4, no. 2, pp.
1087-1095, 1989.
[98] M. Darveniza, “Lightning arrester protection of distribution transformers – revisited”, Proceedings of the 6th International Symposium on
Lightning Protection (SIPDA), São Paulo, pp. 385-389, Nov. 2001.
[99] R. C. Dugan, S. S. Kershaw Jr., and S. D. Smith, “Protecting distribution transformers from low-side current surges”, IEEE Transactions on
Power Delivery, vol. 5, no. 4, pp. 1892-1901, 1990.
[100] M. B. Marz and S. R. Mendis, “Protecting load devices from the effects of low-side surges”, IEEE Transactions on Industry Applications, vol.
29, no. 6, pp. 1196-1203, 1993.
[101] Task Force on Low-Side Surge Requirements for Distribution Transformers, “Secondary (low-side) surges in distribution transformers”, IEEE
[102] J. L. Puri, N. C. Abi-Samra, T. J. Dionise, and D. R. Smith, “Lightning induced failures in distribution transformers”, IEEE Transactions on
Power Delivery, vol. 3, no. 4, pp. 1784-1801, 1988.
[103] G. L. Goedde, R. C. Dugan, and L. D. Rowe, “Full scale lightning surge tests of distribution transformers and secondary systems”, IEEE
[104] R. C. Dugan and S. D. Smith, “Low-voltage-side current-surge phenomena in single-phase distribution transformer systems”, IEEE
[105] P. Hasse, “Surge protective devices (SPDs) – Part I, Requirements of energy coordinated SPDs according to IEC 61312-3”, Proceedings of the
5th International Symposium on Lightning Protection (SIPDA), São Paulo, pp. 130-135, May 1999.
[106] P. Hasse, “Surge protective devices (SPDs) – Part II, Practical application of energy coordinated SPDs according to IEC 61312-3”, Proceedings
of the 5th International Symposium on Lightning Protection (SIPDA), São Paulo, pp. 136-141, May 1999.
[107] F. D. Martzloff and J-S. Lai, “Coordinating current surge protection devices, high-low versus low-high”, IEEE Transactions on Industry
Applications, vol. 29, no. 4, pp. 680-687, 1993.
[108] R. B. Standler, “Coordination of surge arresters and supressors for use on low-voltage mains”, Proceedings of the 9th EMC Zurich, Zurich, pp.
517-524, 1991.
[109] F. D. Martzloff, “Coordination of surge protectors in low-voltage AC power circuits”, IEEE Transactions on Power Apparatus and Systems,
vol. 99, no. 1, pp. 129-133, 1980.
[110] P. Hasse, Overvoltage protection of low voltage systems, IEE Power and Energy Series, no. 33, 2nd edn, The Institution of Electrical
Engineers, 2000.
29

Lightning protection for overhead distribution lines

More Related Content

What's hot (20)

Viewers also liked (18)

Similar to Lightning protection for overhead distribution lines (20)

More from Gilberto Mejía (20)

Recently uploaded (20)

Lightning protection for overhead distribution lines