Field Analysis of the Occurrence of DistributionLine Faults Caused by Lightning Effects

114

IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 53, NO. 1, FEBRUARY 2011

Field Analysis of the Occurrence of Distribution-Line
Faults Caused by Lightning Effects
Teru Miyazaki, Member, IEEE, and Shigemitsu Okabe, Member, IEEE

Abstract—Electric supply reliability is an issue of wide importance to both an information-oriented society and electric power
companies. This paper focuses on lightning effects on distribution
lines. Field research is now underway in the northern part of the
Kanto Plain in Japan, and voltage and current waveforms in distribution lines due to lightning hits have been observed since 1996.
There are now 284 datasets; these include data on 62 direct flashes
to lines. This research reveals that a distribution line can be protected from direct lightning strokes. In some cases, no power follow
current was confirmed after a multiple phase flashover, and a statistical analysis was conducted to investigate factors affecting the
generation of power follow current. These results can serve as a
valuable resource to help clarify the mechanisms underlying the
production of distribution-line faults caused by lightning effects.
Index Terms—Direct lightning stroke, distribution line,
flashover, indirect lightning stroke, lightning protection, power follow current.

I. INTRODUCTION
ISTRIBUTION-line faults can be caused by indirect lightning strokes near distribution lines as well as direct hits
to the line because of their low-insulation level [1]. Electric
power companies have installed lightning-protection devices,
such as shielding wires and surge arresters, to reduce the rate
of distribution-line faults, and these devices can contribute to
higher reliability of the lines than those without the devices [2],
[3]. Experimental studies have pointed out that distribution lines
can be protected from even direct strokes by appropriately arranging surge arresters and shielding wires [4]–[6], due to the
function of those devices to prevent forming the lightning impulse flashover and power-frequency arc [3]. The line fault in

Japanese distribution lines, however, is still mainly caused by
lightning strokes, and how lightning affects lines still remains
a matter for debate. Field studies of lightning-stroke effects,
particularly the mechanisms that produce distribution faults, is
essential for designing rational lightning protection for distribution lines.
Given the aforementioned background, the Tokyo Electric
Power Company has conducted field research of phenomena
accompanying lightning strokes to 6.6 kV distribution systems
since 1996; voltages and currents in the distribution lines were

D

Fig. 1. Construction of distribution lines and observation apparatus. (a) Configuration of current sensors. (b) Measurement points of voltage sensors.

measured directly, and observed lightning effects were classified as direct or indirect using lightning-activated camera systems [7]–[9]. Combining lightning surge waveforms with photographs, provides a comprehensive observation of each lightning flash and its effects. In previous study, each observational
data was introduced in detail [9]. But this paper focuses on
the mechanisms that produce the distribution-line fault, and a
statistical analysis of observed data was conducted aiming to
quantify the factors affecting the production of line faults due
to lightning strokes. The results clarify the effect of lightningprotection devices, such as ZnO elements on the formation of
flashovers or subsequent power follow currents in distribution
lines, and can serve as a valuable resource for designing their
effective placement on distribution lines.
II. OUTLINE OF OBSERVATION SYSTEM
A. Observation Sites and Construction of Distribution Lines

Manuscript received December 17, 2009; revised June 8, 2010; accepted
August 9, 2010. Date of publication December 10, 2010; date of current version
February 16, 2011.
The authors are with the High Voltage and Insulation Group, R&D Center, Tokyo Electric Power Company, Yokohama 230-8510, Japan (e-mail:

; ).
Digital Object Identifier 10.1109/TEMC.2010.2068301

Regions experiencing a high ground-flash density were selected as observation sites in the Kanto Plain [7]–[9]. Field
research has been conducted since 1996.
The constructions of the distribution lines are shown in Fig. 1.
Usually in such regions, shielding wires are installed on the top
of most concrete poles. But in some areas, they are not installed

0018-9375/$26.00 © 2010 IEEE


MIYAZAKI AND OKABE: FIELD ANALYSIS OF THE OCCURRENCE OF DISTRIBUTION-LINE FAULTS CAUSED BY LIGHTNING EFFECTS

115

for observations. Shielding wires are connected to grounding
electrode via common earthing wire by metal wires. Concerning insulation coordination, insulators supporting voltage wires
have the highest insulation level in distribution lines because
their faults affect a power failure of large area. On the other
hand, transformers have the lowest level. But they include ZnO
elements. This means transformers are protected from lightning
strokes.
B. System Configuration
The observation apparatus can simultaneously provide photographs and waveforms for a single lightning event [7]–[9].
In both the camera and the waveform measurement instrument,
the time is recorded using global positioning system (GPS),
which can then be matched with lightning-position and trackingsystem (LPATS) data [10], [11]. Camera systems monitor the
field sites to capture lightning flashes. The cameras are installed
on 63 poles in observation sites and designed to take photos

automatically by controlling the shutter according to external
lightning intensity. Two types of sensors have been used to
observe lightning surge waveforms: a current-measuring sensor and a voltage-measuring sensor. The voltage and current
sensors measure insulator voltage and grounding lead current
waveforms. The measuring sensors are installed on 103 poles
in same sites. Fig. 1 shows an overview of the measurement
points of the sensors. Frequency bandwidth of the sensors is
250 Hz–250 kHz.

Fig. 2. Relation of the number between observed lightning flashes and subsequent faults (1996–2006).
TABLE I
LIGHTNING ATTACHMENT POINT TO DISTRIBUTION SYSTEM. (A) DISTRIBUTION
LINE WITH SHIELDING WIRES. (B) DISTRIBUTION LINE WITHOUT
SHIELDING WIRES

III. OBSERVATIONS
During 11 years from 1996 to 2006, 284 flashes were photographed (62 direct and 222 indirect), with all the observed
events being in summer times. Thus, the discussion in this paper is limited to summer lightning. The percentage of positive
discharges during that period was about 1%.
A. Line Faults by Direct and Indirect Flashes
All of the observed events were classified as direct or indirect
based on the photographs of the flashes. It was noted that direct
denotes a lightning hit to any distribution equipment including
shielding wires, power lines, reinforced concrete poles, and so
on, while a lightning hit to other than the distribution system
was regarded as indirect. Fig. 2 relates the number of direct
and indirect flashes and the subsequent line faults during that
period, causing overcurrent relays in substations to operate. Investigating whether overcurrent relays were operated or not at
the time of the observational photographs GPS time, Fig. 2 data
were collected. These results confirm that direct flashes caused

83% of the faults, which suggests that direct hits were the major
cause of line faults, while the data also show that 48% of all
the direct hits caused faults, whereas only 3% of all the indirect events provided faults. In other words, about half of direct
flashes did not cause line faults, which means that the lines can
be protected from direct hits that were thought would inevitably
damage lines. The authors focused on this issue, and analyzed
observed data from direct hits.

B. Lightning Attachment Point Due to Direct Hits
Based on the photographs of direct flashes, the lightning attachment points were estimated and classified into five patterns.
All results derived are summarized in Table I. Lightning to the
pole head recorded the highest ratio whether shielding wires
were installed or not (see Table (a) and (b)) and the lightning
hits to the power line without shielding wires accounted for only
5%. A report experimentally showed that lightning is likely to
strike bare wires compared to covered wires because the generation of upward leader is suppressed due to insulator [12]. This


116

Fig. 3.

IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 53, NO. 1, FEBRUARY 2011

Direct lightning flash to a pole head without shielding wires.

result suggests that the lightning striking distance of insulated
wires is smaller than that of bare wires. In some cases without shielding wires, as shown in Table I(b), metallic caps were
equipped at the top of reinforced concrete poles, which might
affect the ratio of direct lightning to the power line. In these observational sites, the length of the line equipped with shielding

wires is 10 times as long as that without shielding wires. There
is room for further investigation. Fig. 3 shows an example of
photographs of a lightning flash to a pole head without shielding
wires.
C. Analysis of Observed Direct Lightning Event
Observed surge waveforms due to direct lightning hits, which
are considered to be the primary cause of line faults, were validated using the electromagnetic transients program (EMTP) in
order to estimate whether or not flashover were generated. In
our previous paper [8], one example of direct lightning hit to a
pole head was selected, and a comparison of the observed and
calculated waveforms was conducted. Fig. 4 shows an example
of the observed and calculated insulator voltages, and the details
of the analysis model can be seen in the papers [8], [13]. The
distance between the measurement point and the strike location
is 57 m. The calculated voltage using the flashover model shows
opposite polarity excursions of the waveform starting around 3
μs later when a flashover occurred [see Fig. 4 (b)], which can
also be confirmed in the observation, as shown in Fig. 4(a)
[8]. In this paper, a waveform was also calculated without simulating flashover generation. As shown in Fig. 4 (b), the calculated waveform simulated without flashover generation does
not shows the opposite polarity excursions, which is consistent
with the observed surge waveform, but shows an inversion of
polarity in the wavefront when a flashover was generated [7].

Fig. 4. Comparison between observed and calculated insulator voltage waveforms. (a) Observed waveform. (b) Calculated waveform.

1) Production of Flashover: The production of flashovers
was estimated based on the insulator voltage waveform characteristics described previously.
While a waveform with a rapid reverse of polarity in the
wave front suggests that a flashover must have been produced, a
waveform with the same polarity means no flashover generation.

If the transient is measured far away, e.g., at the end of the
feeder, the numerous reflections of the traveling wave at the
various connections of laterals and transformers are expected to
alter the waveform and make its analysis more difficult.
2) Production of Power Follow Current: The production of
power follow current was also estimated based on the observed
insulator voltage waveforms over a time range up to 20 ms, in
which one cycle of power-frequency (50 Hz) voltages can be
confirmed. Voltages remaining zero over that period at two or
three phases indicate the production of a power follow current,
which will cause an overcurrent relay to operate in a substation.
B. Rate of Flashover

IV. FIELD EXAMINATION OF LINE FAULTS
The production of flashover and power follow current was
examined to clarify the mechanisms that cause distribution-line
faults based on the observed direct-lightning data.
A. Patterns of Lightning Surge Waveforms
Surge voltage waveform characteristics can help estimate
whether flashovers and power follow currents are produced.

Of the observed 62 direct data with cameras, 56 flashes accompanying voltage waveforms were selected as data for the
analysis to estimate whether or not a flashover was produced.
The results are shown in Table II; lightning strokes caused a
flashover to occur in 42 (75%) among 56 cases, and the flashover
in three phases shows the highest percentage (55%). In the case
of flashovers at two phases, flashovers were confirmed at only
R- and T-phases, which is probably because voltage of S-phase
is suppressed compared with the other phases due to coupling

MIYAZAKI AND OKABE: FIELD ANALYSIS OF THE OCCURRENCE OF DISTRIBUTION-LINE FAULTS CAUSED BY LIGHTNING EFFECTS

117

TABLE II
NUMBER OF OBSERVED FLASHOVERS DUE TO DIRECT LIGHTNING HITS

TABLE III
NUMBER OF CONFIRMED POWER FOLLOW CURRENT AFTER FLASHOVERS DUE
TO DIRECT LIGHTNING HITS

Fig. 5. Cumulative frequency distribution of peak current of lightning strokes
based on LPATS data for the observed patterns of flashovers (cumulative probability was calculated for each category).

lower value than after three phases flashover (67%). Even when
power follow currents are generated after three phases flashover
(20 cases), in 13 cases current was generated only at two phases.
In 77% (= 23/30) of the cases that caused flashover to generate at three phases, the arc of the flashover was extinguished at
one phase or more. Taking all the cases into account, the rate
of the generation of power follow currents is a total of 53%.
The fact that only 53%of the data with flashovers caused the
power follow current can provide insight regarding line-fault
mechanism and lead to improving the method of calculation of
distribution outage rate due to lightning, since multiple phase
flashovers are regarded as the fault under present studies, including calculations for line faults due to lightning. The operation of
overcurrent relays is caused by multiple attachment flashovers
and subsequent power follow currents. So “multiple attachment
flashovers” is defined as “faults” practically.
effect. It is noted that 25% of the direct lightning strokes cause no

flashover generation, a remarkable fact given that direct lightning strokes had been thought to inevitably cause flashovers
because of the low-insulation level.
C. Rate of Power Follow Currents
A power follow current after multiple phase flashovers causes
a relay in a substation to operate. Productions of a current on
40 cases were estimated, with associated surge waveforms measured up to 20 ms, in which power follow current can be confirmed, out of 42, the set of direct data with flashover is given in
Table II. Guessing the reason for the different results obtained for
the different conductors, S-phase seems to be the most protected
of the three phases by electromagnetic coupling with shielding
wire because of the shortest length to that. The result is shown
in Table III. No power follow current occurred after flashovers
in single phase. The reason is probably as follows: most distribution lines in Japan use isolated neutral systems. If flashover
occurs only in single phase, the power follow current will be
naturally extinguished due to a small ground-fault current. Neither was a power follow current confirmed after the production
of flashovers in two or three phases, and the rate of power follow current after two phases flashover (20%), in particular, is a

V. FACTORS THAT AFFECT ON LINE FAULTS
In Section IV, the frequencies of the production of flashover
and subsequent power follow current were estimated using
associated surge waveforms. In some cases, multiple phase
flashovers caused no power follow current to generate. It is
important to note that the production of the flashover and power
follow current could be affected by several factors such as shielding wires, surge arresters, impedance of a line, phase angles of
power-frequency voltages, and so on. Further detailed studies
were conducted to evaluate which factors can affect the flashover
phenomena and power follow current due to direct strokes to the
distribution system.
A. Effect of Lightning Current
The amplitude of lightning currents associated with lightning
overvoltages can affect the probability of occurrence of line

faults. Fig. 5 indicates the cumulative frequency of currents amplitude estimated by LPATS data of each flashover types. Fig. 6
indicates the cumulative frequency of power follow currents due
to the observed direct lightning hits. Types of flashovers were
estimated based on associated voltage waveforms. In these figures, each set of distributions was plotted by selecting lightning
currents associated with particular events in terms of types of


118

IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 53, NO. 1, FEBRUARY 2011

Fig. 6. Cumulative frequency distribution of peak current of lightning strokes
based on LPATS data for the observed patterns of power follow current after flashover at three phases (cumulative probability was calculated for each
category).

flashover or subsequent power follow current. There were no
data exceeding 60 kA, and these distributions show the median 25 kA, showing almost no characteristic difference from
previous studies [14], [15]. In Fig. 5, the cases of single-phase
flashovers are only six. So they should be collected and analyzed
to find the correlation lightning currents, peak of transient voltage clearly in future. A few remarks should be made concerning
uncertainness of LPATS data. An experimental study suggests
that negative first stroke currents estimated by LPATS tend to be
smaller than first peaks of measured values especially for large
currents exceeding 40 kA [10], [11]. When looking at Figs. 4
and 5, the lightning currents amplitudes inferred by LPATS
system could have underestimate the actual values. A study reported that median variability of lightning currents inferred by
the lightning location system (LLS) was estimated to be from
20% to 30% [16]. The “20%–30%” uncertainty in peak-current
estimates in [16] is for negative subsequent strokes in existing
channels. The uncertainty can only be larger for first strokes.

The accuracy of lightning currents estimated by this LLS is estimated to be equivalent to that by the LPATS in our research.
Thus, it is considered that Figs. 4 and 5 can be evaluated with
the aforementioned accuracy.

Fig. 7. Relation of the number between observed direct lightning strokes and
multiple phase flashovers and line faults.

C. Effect of Phase Angle of Power-Frequency Voltage
Phase angles of the power-frequency voltage when the
flashover occurred was estimated based on the observed surge
waveforms due to direct lightning hits. The phase angles were
confirmed in the waveforms of 52 events of direct lightning,
and Fig. 8 shows the phase angles of the line voltage when
two or three phase flashovers were generated, showing almost
no characteristic having effects on the production of flashovers.
Fig. 9 also shows the phase angles of line voltage when the subsequent power follow current was generated after flashovers at
three phases. When the power follow currents were generated at
two phases, most angles concentrate around the maximum value
of the power-frequency voltage [see Fig. 9(a)]. These facts suggest that the phase angle of power-frequency voltage can have
an effect on the production of the power follow current, which
results in distribution-line faults. In the case of the power follow
current at three phases, the phase angles are dispersed and no
relation between the angle and the power follow current was
confirmed, as shown in Fig. 9(b).

B. Effect of Shielding Wires

D. Effect of Impedance Between a Substation and Strike
Location

Fig. 7 relates the number of direct lightning strokes, subsequent multiple phase flashovers and line faults to the presence
or absence of the shielding wire on the struck points. When the
shielding wire was installed above the lines, the proportion of
the events of the multiple phase flashovers is 56% (= 22/39),
which is low compared to the value of 82% (= 14/17) in the
case with no shielding wire on the struck point.
On the other hand, the proportion of line faults after multiple phase flashovers in the case with shielding wire is 59%
(= 13/22), which is almost the same as the value of 57% (=
8/14) in the case with no shielding wire. These suggest that the
presence of shielding wires on the struck point plays a more important role in the production of the multiple phase flashovers
than that of the subsequent power follow current.

The impedance of the lines between a substation and attachment point was investigated because that can affect the shortcircuit current during the duration of the power follow current.
The impedance was calculated by cable length from a substation
to struck point using impedance value per 1 m shown in specifications. Fig. 10 shows the cumulative frequency distribution
of the impedance in the cases with or without flashover, and the
impedance was calculated on the basis of 10 MVA. The shape
of both distributions was roughly the same. Fig. 11 shows the
cumulative frequency distribution of impedance to the patterns
of the power follow current after flashovers at three phases, and
each shows the same distribution. The correlation between the
impedance and the production of the flashover or power follow
current is estimated to be weak in the lines.


MIYAZAKI AND OKABE: FIELD ANALYSIS OF THE OCCURRENCE OF DISTRIBUTION-LINE FAULTS CAUSED BY LIGHTNING EFFECTS

Fig. 8. Phase angles of line voltage when flashover was generated (data were
plotted for each category). (a) Flashover at two phases. (b) Flashover at three
phases.

119

Fig. 9. Phase angles of line voltage when power follow current was generated
after three phase flashovers (data were plotted for each category). (a) Power
follow current at two phases. (b) Power follow current at three phases.

E. Effect of the Distance Between ZnO Elements and the Strike
Location
Next analysis was the distance between a pole with ZnO
elements, such as a surge arrestor and a struck point. Fig. 12
shows the cumulative frequency distribution of the distance
between the ZnO elements and the struck point. The cases
without flashovers represent lower distances than those with
flashovers. If looking at the 50% value, it is 64 m in the case
of flashover while 30 m without flashover. Fig. 13 shows the
cumulative frequency distribution of the distance to the patterns
of the subsequent power follow current after the production of
flashover at three phases. The distances show lower values in
case of no power follow current compared to cases with the
production of the power follow current, and the distances in
the occurrence of the current at three phases show higher value
than those at two phases. These results reveal that the distance
between the ZnO elements and the strike point plays a decisive
role in protecting the lines against direct lightning strokes, since
it affects both the formation of flashover and power follow
current. But the measured distance in Figs. 11 and 12 seems
to include expected errors, 10 m at the most, by measurement
from two or more directions.
F. Evaluation of the Results

It may be desirable to briefly review the results. In Section V,
several factors, such as the effects of ZnO elements were evalu-

Fig. 10. Cumulative frequency distribution of the impedance of the lines
between a substation and a struck point in the cases with or without flashover
(cumulative probability was calculated for each category).

ated for their potential impact on the production of the flashover
and power follow current due to direct lightning strokes. The
results are summarized in Table IV.
1) Generation of Flashover: The observation confirms two
factors that effect the production of flashovers: the presence or
absence of shielding wires on the struck pole, and the distance
between the ZnO elements and the struck point.
2) Generation of Power Follow Current: The distance between the ZnO elements and the strike point also affect the formation of the power follow current as well as flashover, while
the effect of the phase angle of power-frequency voltage was


120

IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 53, NO. 1, FEBRUARY 2011

TABLE IV
ESTIMATION OF FACTORS THAT AFFECT THE PRODUTION OF FLASHOVER AND
POWER FOLLOW CURRENT

Fig. 11. Cumulative frequency distribution of the impedance of the lines
between a substation and a struck point to the patterns of the power follow
current after flashovers at three phases (cumulative probability was calculated
for each category).

Fig. 12. Cumulative frequency distribution of the distance between ZnO elements and a struck point in the cases with or without flashover (cumulative
probability was calculated for each category).

Fig. 13. Cumulative frequency distribution of the distance between ZnO
elements and a struck point to the patterns of the power follow current after flashovers at three phases (cumulative probability was calculated for each
category).

confirmed in the production of power follow current only at two
phases after three phases flashovers.
VI. LIGHTNING-PROTECTION DEVICES IN
DISTRIBUTION SYSTEMS
The analysis of the observation indicates that ZnO elements
have a function to prevent forming flashovers and subsequent
power follow currents at poles even without ZnO elements. Let

us devote a little more space to consider it from viewpoint of the
traveling of surge waveforms. A lightning surge travels through
common earth wire (or shielding wire) from the injection point
to the closest ZnO elements, and the current flows into grounding
electrode. Then, the insulator voltages of a struck pole will be
suppressed after the reflection of the surge voltage travels back
to the struck pole. Thus, the effects of ZnO elements of another
pole reaches the struck pole with time delay, and this might be
one of the causes why the distance between ZnO elements and a
struck point affects the formation of the power follow current as
well as the flashover. Schoene et al. [17] obtained experimental data from test with rocket-triggered lightning currents, and
estimated that arresters reduce surge voltages against a direct
lightning stroke with time delay. On the other hand, shielding
wires have an effect on the generation of only flashover. The

explanation for this is probably that the effect of shielding wires
on the reduction of insulator voltages can be obtained immediately after lightning hit to a pole because of the diversion into
the wires. Let us, for the moment, consider other factors. The
results also confirmed no effect of the lightning current or the
impedance between a substation and a struck point. With regard
to the production of the power follow current, the correlation of
phase angles of power-frequency voltage was confirmed only
after the case of the production of the power follow current at
two phases after the three phases flashovers. The facts indicate
that the lightning-protection devices, such as shielding wires
and ZnO elements play a more decisive role against lightning
stokes compared to the effect of other factors in the lines when
considering distribution systems with a number of lightningprotection devices. But excessive arresters (e. g., on every pole)
do not always protect distribution lines. A consideration has to
be given to the arrester failures, which will cause more serious
outage rate of lines than ordinary flashovers [18]. Regarding the
effect of arrestors in the observed areas, the authors also confirmed that most insulator voltages tend to concentrate less than
36 kV, which is the limit voltage of ZnO elements in a transformer [8]. These results indicate that surge voltages do not
grow in proportion to the lightning-stroke current, and which
can be one of the causes why lightning currents seem to have
no effect on the formation of line faults as explained in this


MIYAZAKI AND OKABE: FIELD ANALYSIS OF THE OCCURRENCE OF DISTRIBUTION-LINE FAULTS CAUSED BY LIGHTNING EFFECTS

section. It is noted that a correlation between insulator voltages
and lightning-stroke currents estimated by LPATS was investigated in our previous study [9], and the analysis also confirmed
that the distance between a ZnO elements and a strike location
affects surge voltages due to lightning strokes.
VII. CONCLUSION

This paper mainly focuses on the observed direct lightning
events to the lines, and analyses were conducted to clarify which
factors can affect the production of the flashover and power
follow current due to direct strokes.
The results mainly confirm two factors that affect the production of flashovers: the presence or absence of shielding wires on
the struck point, and the distance between ZnO elements and a
struck point. The distance between the ZnO elements and the
strike location is estimated to have affected the generation of the
power follow current as well as flashover, and the effect of the
phase angle of power-frequency voltage was confirmed in the
generation of power follow current only at two phases after three
phases flashovers, while no relation between the phase angle and
the power follow current was confirmed in the other cases.
It should be concluded, from what has been said previously,
that several factors can affect the generation of line faults and
lightning-protection devices, such as surge arrestors and shielding wires are effective for the protection of distribution lines
against direct lightning strokes.
In future, the authors will work on designing rational lightning
protection for distribution lines by analyzing more observation
data and improving a method for the calculation of distributionline faults.
VIII. ACKNOWLEDGMENT
The authors would like to thank S. Amemiya of Tokyo Electric Corporation, who extended support and cooperation to this
study. They are also grateful for the support and cooperation of
related departments, such as the distribution department of their
company.
REFERENCES
[1] P. D. Kannu and M. J. Thomas, “Lightning induced voltages on multiconductor power distribution line,” Inst. Elect. Eng. (IEE) Proc. Gener.
Transm. Distrib., vol. 152, no. 6, pp. 855–863, Nov. 2005.
[2] 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 Trans. Power Del., vol. 19,
no. 1, pp. 423–431, Jan. 2004.
[3] J. He, S. Gu, S. Chen, R. Zeng, and W. Chen, “Discussion on measures
against lightning breakage of covered conductors on distribution lines,”
IEEE Trans. Power Del., vol. 23, no. 2, pp. 693–702, Apr. 2008.
[4] S. Yokoyama, “Lightning protection of overhead power distribution lines,”
Inst. Elect. Eng. Jpn. (IEEJ) Trans. Power Del., vol. 22, no. 4, pp. 2236–
2244, 2007.
[5] P. Barker, “Photography helps solve distribution lightning problems,”
IEEE Power Eng. Rev., vol. 13, no. 6, pp. 23–26, Jun. 1993.
[6] H. Taniguchi, H. Sugimoto, and S. Yokoyama, “Observation of lightning
performance on power distribution line by still camera,” in Proc. 23rd
Int. Conf. Lightning Protection, 1996, vol. 1, pp. 119–124.
[7] T. Takao, S. Okabe, T. Miyazaki, and K. Aiba, “Analysis of lightning
phenomena observed in distribution lines,” in Proc 28th Int. Conf. on
Lightning Protection, no. VI-2, Sep. 2006.

121

[8] T. Miyazaki, S. Okabe, and S. Sekioka, “An experimental validation of
lightning performance in distribution lines,” IEEE Trans. Power Del.,
vol. 23, no. 4, pp. 2182–2190, Oct. 2008.
[9] T. Miyazaki and S. Okabe, “Detailed field study of lightning stroke effects
on distribution lines,” IEEE Trans. Power Del., vol. 24, no. 1, pp. 352–359,
Jan. 2009.
[10] T. Shioda, T. Narita, E. Zaima, and M. Ishii, “Performance evaluation of
LPATS-T at TEPCO,” in Proc. 25th Int. Conf. on Lightning Protection,
2000, pp. 170–175.
[11] T. Shioda, N. Fukiyama, A. Mochizuki, E. Zaima, M. Ishii, and K. Cummins, “Performance evaluation of new generation LPATS at TEPCO,” in

Proc. 24th Int. Conf. on Lightning Protection, 1998, pp. 162–167.
[12] Y. Hashimoto, S. Yokoyama, T. Yokota, and A. Asakawa, “Studies on
characteristics of lightning stroke distance to power distribution linesdischarge characteristics of open wire and insulated wire,” Trans Inst.
Elect. Eng., Jpn, vol. 115-B, no. 12, pp. 1508–1514, Dec. 1996.
[13] S. Sekioka, “Lightning-surge analysis model of reinforced concrete pole
and grounding lead conductor in distribution line,” presented at the Int.
Workshop High Voltage Eng., Sapporo, Japan, 2004.
[14] R. B. Anderson and A. J. Eriksson, “Lightning parameters for engineering
Application,” Int. Conf. on Large Electric High-Tension Systems (CIGRE)
Electra, no. 69, pp. 65–102, Mar. 1980.
[15] K. Berger, R. B. Anderson, and H Kroeninger, “Parameters of Lightning
flashes,” Int. Conf. on Large Electric High-Tension Systems (CIGRE)
Electra, vol. 41, pp. 23–27, Jul., 1975.
[16] K. L. Cummins, M. J. Murphy, E. A. Bardo, W. L. Hiscox, R. B. Pyle,
and A. E. Piper, “A combined TOA/MDF technology upgrade of the U. S.
national lightning detection network,” J. Geophysical Research, vol. 103,
no. D8, pp. 9035–9044.
[17] J. Schoene, M. Uman, V. Rakov, A. Mata, C. Mata, K. Rambo, J. Jerauld,
D. Jordan, and G. Schnetzer, “Direct lightning strikes to test power distribution lines—Part I: Experiment and overall results,” IEEE Trans. Power
Del., vol. 22, no. 4, pp. 2236–2244, Oct. 2007.
[18] T. E. McDermot, T. A. Short, and J. G. Anderson, “Lightning protection
of distribution lines,” IEEE Trans. Power Del., vol. 9, no. 1, pp. 138–152,
Jan. 1994.

Teru Miyazaki (M’07) received the B.Eng. and
M.Eng. degrees in electrical engineering from
the University of Electro-Communications, Tokyo,
Japan, in 1995 and 1997, respectively. He received
the Doctorate degree from the Shonan Institute of
Technology, Kanagawa, Japan, in 2008.

In 1997, he joined Tokyo Electric Power Company, Tokyo, Japan, where he is currently a member
of the High Voltage & Insulation Group, the R & D
center. His main research interest includes the lightning protection design of a power distribution line.

Shigemitsu Okabe (M’98) received the B.Eng.,
M.Eng., and Dr. degrees in electrical engineering
from the University of Tokyo, Tokyo, Japan, in 1981,
1983, and 1986, respectively.
He has been with Tokyo Electric Power Company
since 1986, where he is currently the Group Manager
of the High Voltage & Insulation Group, the R &
D Center. In 1992, he was a visiting scientist at the
Technical University of Munich. He has also been
a Guest Professor at the Doshisha University since
2005, at the Nagoya University since 2006, and a
Visiting Lecturer at the Tokyo University. He is a Secretary/Member at several
WG/MT in CIGRE and IEC.
Dr. Okabe is an Associate Editor of the IEEE Dielectrics and Electrical
Insulation.