See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/256292931

Protection of distribution transformer against arising or transferred fast-

front overvoltages: Effects of surge arrester connection conductors length

Conference Paper · August 2013

CITATIONS READS

6 5,301

5 authors, including:

Zacharias G. Datsios Pantelis N. Mikropoulos

Aristotle University of Thessaloniki Aristotle University of Thessaloniki
36 PUBLICATIONS   302 CITATIONS    141 PUBLICATIONS   1,357 CITATIONS   

SEE PROFILE SEE PROFILE

Thomas E. Tsovilis
Aristotle University of Thessaloniki
86 PUBLICATIONS   628 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Soil ionization View project

Surge Protection View project

All content following this page was uploaded by Zacharias G. Datsios on 16 May 2014.

The user has requested enhancement of the downloaded file.

 OB2-04

PROTECTION OF DISTRIBUTION TRANSFORMER AGAINST ARISING

OR TRANSFERRED FAST-FRONT OVERVOLTAGES: EFFECTS OF
SURGE ARRESTER CONNECTION CONDUCTORS LENGTH
Z. G. Datsios1*, P. N. Mikropoulos1, Z. Politis2, A. G. Kagiannas2 and T. E. Tsovilis2
1
High Voltage Laboratory, School of Electrical & Computer Engineering,
Faculty of Engineering, Aristotle University of Thessaloniki, 54124 Thessaloniki, Greece
2
Raycap Corporation, 14 Telou and Petroutsou, 15124 Athens, Greece
*Email: zdatsios@auth.gr

Abstract: The effects of the length of the surge arrester connection conductors on the
lightning surges impinging on a typical wood pole-mounted 50 kVA, 20/0.4 kV transformer
of the Hellenic distribution system are investigated through detailed ATP-EMTP
simulations. By considering both first and subsequent direct lightning strokes to a
connected overhead distribution line, the effectiveness of the common practice
transformer protection scheme and of an alternative one utilizing shorter surge arrester
connection conductors in suppressing fast-front overvoltages was evaluated. A shorter
length of the surge arrester connection conductors results in a reduction in the amplitude
of the overvoltages arising at the medium-voltage terminals of the transformer and in a
slower rate of increase of the overvoltage amplitude with lightning return-stroke current.
The overvoltages transferred to the low-voltage terminals of the transformer are
practically not affected by the length of the surge arrester connection conductors.
Protection against transferred overvoltages was provided by surge protective devices
installed at the low-voltage terminals of the transformer. By utilizing shorter surge arrester
connection conductors the transformer failure rate, estimated through risk assessment, is
reduced by approximately 11%.

1 INTRODUCTION accurate evaluation of the protection afforded by

surge arresters to the transformer can be made
The fast-front overvoltages arising at distribution with the aid of detailed simulations using an
equipment utilizing non-self-restoring insulation, electromagnetic transient analysis program.
such as transformers or cables, may cause
permanent failure resulting in system outages and In this study the effects of the length of the surge
economic losses. Therefore, distribution equipment arrester connection conductors on the fast-front
is most commonly protected against impinging overvoltages arising at a typical pole-mounted
lightning surges by surge arresters and surge 50 kVA, 20/0.4 kV transformer of the Hellenic
protective devices. It is well known that the distribution system are investigated through
efficiency of the afforded protection is affected by detailed ATP-EMTP [9] simulations. By considering
the length of the connection conductors; the latter both first and subsequent direct lightning strokes to
should be as short as possible in order to achieve the connected medium-voltage overhead line, the
optimum protection [1-6]. common practice transformer protection scheme
and an alternative one utilizing shorter surge
According to common practice, distribution arrester connection conductors were evaluated. A
transformers are protected against impinging better lightning performance of the transformer is
lightning surges by surge arresters installed close achieved by implementing the alternative than the
to their medium-voltage bushings. It can be shown common practice protection scheme; the
theoretically that the fast-front overvoltages arising transformer failure rate, estimated through risk
at a protected transformer increase in amplitude assessment, is reduced by approximately 11%.
with the steepness of the incoming lightning surge
and that the greater the separation distance 2 MODELLING OF THE EVALUATED
between surge arresters and transformer the less SYSTEM
is the effectiveness of the provided protection.
Actually, analytical methods for the estimation of Figure 1a shows the common practice protection
the lightning overvoltages arising at a protected scheme of a typical wood pole-mounted 50 kVA,
transformer can be found in [7, 8]. However, as 20/0.4 kV transformer of the Hellenic distribution
theoretical analysis is based on several system. An alternative configuration is shown in
simplifications, analytical results should be Figure 1b; surge arresters are mounted on an
considered as conservative yet acceptable additional wood crossarm, considerably closer to
estimates for the protective distance and the safety the transformer terminal bushings. The total length
margin provided by surge arresters. A more of the surge arrester connection conductors is 6 m

303
 OB2-04

and 1.1 m for the common practice and the the line positioned at the greatest distance from the
alternative protection scheme, respectively. It must middle phase conductor (Figure 3), at the time
be noted that, in accordance with [5, 10], surge instant of negative power-frequency voltage peak
durable fuses coordinated with the surge arresters of the struck phase and at a distance of 100 m
should be used for the alternative protection from the wood pole-mounted substation. This
scheme. Both configurations were employed in distance corresponds to the maximum span length
detailed ATP-EMTP [9] simulations to assess the along rural overhead lines of the Hellenic
effect of the surge arrester connection conductors distribution system. Both first and subsequent
length on the fast-front overvoltages arising at the lightning return-strokes to the overhead line were
transformer. Figure 2 shows the schematic considered. Lightning stroke was represented by a
diagram of the system employed in simulations. A current source producing a current with front
medium-voltage (MV) overhead line terminates at upwardly concave [11]. According to CIGRE [11],
a 20/0.4 kV transformer. The transformer, sharing the median values of the front time and maximum
the same grounding system with the surge current steepness are a function of the lightning
arresters, feeds through a short low-voltage (LV) first return-stroke peak current; the median of the
overhead service line a 3-phase symmetrical load; time to half value is equal to 77.5 μs. For the
the latter uses a separate grounding system. subsequent return-stroke current, the median
values of the front time and time to half value are
Simulations were performed by assuming negative equal to 0.67 μs and 30.2 μs, respectively, and the
lightning flashes to the MV overhead line; as a median value of the maximum current steepness is
worst case scenario, the overhead line terminates a function of the peak current [11].
at the transformer (dead-end configuration).
Lightning strikes to the outer phase conductor of The MV overhead line (Figure 3) and the LV
service line were represented by J.Marti frequency-
dependent models considering line geometry. Line
parameters were calculated for a soil resistivity of
200 Ωm. The distribution transformer (50 kVA,
20/0.4 kV, Dyn1, 4%) was modelled by a
capacitance π-circuit together with a BCTRAN
model. The 3-phase symmetrical load connected to
the 20 m long LV line was simulated according to
[12]. Surge arresters were represented by a
frequency-dependent model [13] (Figure 4), with
parameters calculated based on the surge arrester
characteristics given in Table 1. Surge protective
devices (SPD) were modelled as nonlinear current
dependent resistors taking into account their

Figure 1: Wood pole-mounted 50 kVA, 20/0.4 kV

distribution substation; protection schemes of the
transformer: (a) common practice in the Hellenic
distribution system, (b) alternative configuration

Figure 3: Overhead line (20 kV) of the Hellenic

distribution system; wood crossarms support the
porcelain pin-type insulators (BIL=125 kV)

Figure 4: Frequency-dependent surge arrester

Figure 2: Schematic diagram of the evaluated model [13]; parameters calculated based on the
system surge arrester characteristics given in Table 1

304
 OB2-04

Table 1: Surge arrester (SA) and surge protective

device (SPD) characteristics
SA SPD
Rated1 and Nominal2 voltage: 1
21 kV 2
240 V
Continuous operating voltage: 16.8 kV 300 V
Nominal discharge current: 5 kA, 8/20 μs 10 kA, 8/20 μs
High current impulse: 65 kA, 4/10 μs 200 kA, 8/20 μs
Residual voltage for nominal
58 kV 800 V
discharge current:

characteristics (Table 1). Connection conductors

with lengths shorter than 3 m were modelled as
lumped parameter inductances of 1 μH/m [14].
Lengths longer than 3 m were modelled as
lossless distributed parameters line segments with
surge impedance of 400 Ω. Flashover of line
insulation was modelled by employing the
integration method [1, 3, 15], which was
implemented in ATP-EMTP by the ISF object [16].
The latter is connected between phase conductors, Figure 5: Typical waveshapes of overvoltages
since for the particular line (Figure 3) the flashover arising at the MV terminals of the transformer due
path with the lowest critical flashover voltage to lightning (a) first and (b) subsequent strokes to
(CFO) is between phase conductors along the the MV overhead line; 9 kA lightning peak current
wood crossarm. The parameters of the integration
method were determined according to [17, 18] voltage reflections due to the different surge
considering the total CFO of line insulation; the impedances between the line and the transformer.
latter was estimated respectively as 482 kV and This oscillation has a higher frequency and occurs
357 kV between left and middle phase conductors earlier in a faster rising wavefront of the arising
and right and middle phase conductors (Figure 3), overvoltage for subsequent strokes due to the
by using the extended CFO-added method [19]. higher steepness of the subsequent return stroke
The arc channel following flashover of line current. The effect of such oscillations, occurring
insulation was represented by an inductance of around the peak of an impulse overvoltage, on the
1 μH/m [20] for simplicity. dielectric behaviour of insulation is frequency
dependent. The peak value of the oscillation
The concentrated grounding systems of the pole- affects gradually lesser the dielectric strength with
mounted substation and the load were modelled as increasing frequency; this effect becomes
current-dependent resistances, considering thus, negligible for frequencies higher than
the decrease of the grounding impedance to approximately 3 MHz [24]. Thus, in order to
values lower than the initial low current and low compare the arising overvoltages with the basic
frequency grounding resistance caused by soil insulation level (BIL) of the transformer, the
ionization. Weck’s simplified grounding system “equivalent” overvoltage that the insulation is
model, adopted by CIGRE [11], was employed in subjected to under lightning impulse voltage was
simulations, implemented in ATP-EMTP by the estimated by adopting the procedure for manual
TGIR object [21]. The low current and low calculation of impulse parameters from graphical
frequency grounding resistance was taken 10 Ω for waveforms according to [24].
the load and 100 Ω for the substation. Such high Figure 6 shows the variation of the amplitude of the
transformer grounding resistance values, resulting highest “equivalent” phase-to-phase, UMeq, and
in more severe stress for the LV side of the phase-to-neutral, VMeq, overvoltages at the MV
transformer [22], have been reported in the terminals of the transformer with lightning peak
Hellenic distribution system and associated with current for both first and subsequent lightning
transformer failures due to lightning [23]. strokes. Obviously UMeq is significantly higher than
the VMeq. In addition, both overvoltages increase
3 SIMULATION RESULTS AND DISCUSSION with increasing lightning peak current however with
a different rate of increase depending on the length
3.1 Overvoltages arising at the MV terminals of the surge arrester connection conductors. A
of the distribution transformer shorter length results in lower overvoltage
Figure 5 shows typical waveshapes of the amplitudes (up to 27%), especially with increasing
overvoltages arising at the MV terminals of the lightning return-stroke current, and in a slower rate
transformer due to lightning first and subsequent of increase of the overvoltage amplitude with
strokes to the MV overhead distribution line. There lightning peak current. These effects are more
is a damped oscillation superimposed on the pronounced for subsequent than first lightning
arising fast-front overvoltages, associated with strokes to the MV overhead distribution line.

305
 OB2-04

180

160
Equivalent overvoltage peak, kV

140
BIL=125 kV
120

100

80

60
UMeq common practice
40
UMeq alternative
20 VMeq common practice
VMeq alternative
0
2 4 6 8 10 12 14 16 18
Lightning peak current, kA
Figure 6: Equivalent overvoltages at the MV
terminals of the distribution transformer as a
function of lightning first and subsequent return-
stroke peak current, depicted respectively by
continuous and dashed lines
For the common practice protection scheme of the Figure 7: Typical waveshapes of overvoltages
transformer, the “equivalent” overvoltage exceeds transferred to the LV terminals of the transformer
the corresponding BIL (125 kV) for direct lightning due to lightning (a) first and (b) subsequent strokes
flashes to the MV line with first and subsequent to the MV overhead line; 9 kA lightning peak
return-stroke current higher than about 8 kA and current, SPD not installed
6 kA, respectively (Figure 6). It is important that for
80
the case of the alternative configuration protection VLeq common practice
VLeq alternative
scheme the corresponding limiting currents are 70
Equivalent overvoltage peak, kV

higher, about 17 kA and 12.5 kA for first and 60

subsequent return-strokes, respectively.
50

3.2 Overvoltages transferred to the LV 40

terminals of the distribution transformer
BIL=30 kV
30
Transferred overvoltages to the LV side of the
transformer are mainly associated with the 20

potential rise of the transformer grounding due to 10 with SPD

the current flowing through the MV surge arresters

0
and to the electromagnetic coupling between MV 2 4 6 8 10 12 14 16 18
Lightning peak current, kA
and LV transformer terminals. Figure 7 shows
typical waveshapes of the overvoltages transferred Figure 8: Equivalent overvoltages at the LV
to the LV side of the transformer when SPD are not terminals of the distribution transformer as a
installed at the LV terminals. There is a function of lightning first and subsequent return-
superimposed damped oscillation associated with stroke peak current, depicted respectively by
voltage reflections due to the different surge continuous and dashed lines
impedances between transformer and LV service
line. Thus, in order to compare the transferred the BIL (30 kV) of the LV side of the transformer for
overvoltages with the BIL of the LV side of the lightning flashes to the MV overhead distribution
transformer (30 kV), the “equivalent” overvoltage line with first and subsequent return-stroke
that the insulation is subjected to under lightning currents higher than about 5.5 kA and 4.5 kA,
impulse voltage, was estimated according to [24]. respectively. It is important, however, that the
lightning stroke peak current causing a LV side
Figure 8 shows the variation of the computed transformer failure depends on load and
amplitude of the highest “equivalent” phase-to- transformer grounding resistances, increasing as
neutral overvoltage, VLeq, transferred to the LV the latter decreases [22]. Nevertheless, for the
terminals of the transformer with lightning current. evaluated system there is a need for protection
The “equivalent” phase-to-phase overvoltages against transferred overvoltages by installing SPD
were omitted from this graph as they were found at the LV terminals of the transformer. In this case,
significantly lower than both VLeq and BIL of the LV the transferred overvoltages are greatly reduced to
side of the transformer. From Figure 8 it is obvious values significantly lower than the BIL of the LV
that VLeq, increasing with lightning peak current, is side of transformer (Figure 8). The amplitude of the
practically not affected by the reduction of the transferred overvoltage, being primarily dependent
length of the surge arrester connection conductors. on the characteristics of the SPD, is practically not
It is must be noted that for both protection affected by variations in lightning return-stroke
schemes when SPD are not installed VLeq exceeds current.

306
 OB2-04

4 RISK ASSESSMENT Table 2: Lightning stroke parameters causing

transformer failure
The failure rate of a distribution transformer, R Common Alternative
(failures/yr), due to direct lightning flashes to practice configuration
incoming overhead lines can be estimated as the IF, kA 8 17
sum of the failure rates due to lightning first and IS, kA 6 12.5
subsequent strokes, RF (failures/yr) and RS P(I>IF) 0.959 0.774
(failures/yr), respectively. By assuming no P(I>IS) 0.912 0.488
correlation between the first and subsequent
return-stroke current amplitudes, RF and RS can be
probabilities of the prospective lightning return-
expressed as
stroke current being greater than these values,
RF  n  N  LS  P I  IF  (1a) calculated according to (3) and (4). Figure 9 shows
the estimated failure rate of the distribution
RS  n  N  LS  1  P I  IF    P  I  IS  (1b) transformer due to direct lightning flashes to the
MV overhead distribution line as a function of the
where n is the number of the lines connected to the limit distance. It is evident that the increase in
substation and LS is the limit distance from the failure rate with limit distance is more pronounced
substation along the overhead line within which a for the common practice than the alternative
lightning flash may cause failure to the transformer, configuration protection scheme. Also, from Figure
that is, an incoming surge exceeding the BIL of the 9 it can be deduced that the reduction in the
transformer. N (flashes/100km/yr) is the annual transformer failure rate due to the shorter surge
number of lightning flashes to an overhead line; for arrester connection conductors is about 11%. As a
lines in open ground N can be estimated by (2) [19]. rough approximation, for the evaluated system the
transformer failure rate per km of limit distance due
N  0.1N g  28h0.6  b  (2) to direct lightning flashes to the incoming overhead
line can be expressed as 0.1Ng and 0.09Ng for the
In (2) Ng (flashes/km2/yr) is the ground flash common practice and alternative configuration,
density, h (m) is the height of the uppermost respectively. It is noteworthy that the contribution
conductor at the pole and b (m) is the separation of subsequent strokes to the total transformer
distance between the outer phase conductors. It failure rate was found less than 12%.
must be mentioned however that the estimation of
N would depend on the lightning attachment model It must be mentioned that as simulations were
adopted for the evaluation of the lightning performed without considering corona damping
performance of the overhead distribution line. effects on the lightning surges propagating along
P(I>IF) and P(I>IS) are the probabilities of the the overhead distribution line, the transformer
prospective lightning return-stroke current being failure rate shown in Figure 9 should be considered
greater than IF and IS, respectively. IF and IS (kA) as upper limit. Moreover, the risk assessment
are respectively the minimum first and subsequent results would depend on the lightning peak current
return-stroke peak currents of all possible lightning distribution, which varies seasonally and
flashes that may terminate within LS causing a geographically, and on the extent of shielding
failure to the transformer. P(I>IF) and P(I>IS) can against direct lightning strokes provided by nearby
be estimated as structures in the region which the distribution line
  crosses [26].
P  I  IF    f  I dI , P  I  IS    f  I dI (3)
IF IS Common practice
1.0 Alternative configuration
where f(I) is the log-normal probability density
function of the lightning peak current distribution 0.8
Failure rate, failures/yr

given as [11, 19, 25]

 ln I  ln I 
0.6
2

1
f I   exp   (4)
2 lnI  2 ln2  0.4
 
where Ī, σln are respectively the median value and 0.2

the standard deviation of the natural logarithm of

the lightning peak current distribution. According to 0.0
[25], Ī=30.1 kA and σln=0.76 for first and Ī=12.3 kA 0 400 800 1200 1600 2000 2400 2800
Limit distance, m
and σln=0.5296 for subsequent return-strokes. Figure 9: Failure rate of distribution transformer,
Table 2 summarises the minimum first and R, due to direct lightning first and subsequent
subsequent return-stroke peak currents causing strokes to the MV overhead distribution line as a
failure to the distribution transformer, derived from function of limit distance, Ls; Ng = 4 flashes/km2/yr,
Figure 6. This table also shows the corresponding line height 8 m

307
 OB2-04

5 CONCLUSIONS voltage and low voltage networks against lightning.

Part 2: Lightning protection of medium voltage
Detailed ATP-EMTP simulations have been networks”, Tech. Bul. 441, 2010
performed to investigate the effect of the length of [9] Canadian-American EMTP Users Group: “ATP Rule
the surge arrester connection conductors upon Book,” 1997
impinging lightning surges on a pole-mounted [10] C. L. Smallwood, H. S. Regina and C. J. Cook:
20/0.4 kV distribution transformer. Both first and “Nuisance operations of distribution fuse links due to
lightning-induced current surges”, IEEE Trans. Ind.
subsequent direct lightning strokes to a connected Appl., vol. 43, no. 1, pp. 196-201, 2007
MV overhead line were considered. A shorter
[11] CIGRE WG 33.01: “Guide to procedures for
length of the surge arrester connection conductors, estimating the lightning performance of transmission
utilized in an alternative to the common practice lines”, Tech. Bul. 63, 1991
transformer protection scheme, results in a [12] H. K. Hoidalen: “Lightning-induced voltages in low-
reduction in the amplitude of the overvoltages voltage systems and its dependency on voltage line
arising at the MV terminals of the transformer and terminations”, in Proc. 23rd ICLP, Birmingham, U.K.,
in a slower rate of increase of the overvoltage 1998, pp. 287-292
amplitude with lightning return-stroke current. [13] P. Pinceti and M. Giannettoni: “A simplified model
These effects are more pronounced for for zinc oxide surge arresters”, IEEE Trans. Power
subsequent than first lightning direct strokes to the Del., vol. 14, no. 2, pp. 393-398, 1999
overhead distribution line. The overvoltages [14] IEEE Task Force: “Modeling guidelines for fast front
transferred to the LV terminals of the transformer transients”, IEEE Trans. Power Del., vol. 11, no. 1,
are practically not affected by the length of the pp. 493-506, 1996
surge arrester connection conductors. [15] M. Darveniza and A. E. Vlastos: “The generalized
integration method for predicting impulse volt-time
By implementing the alternative than the common characteristics for non-standard wave shapes-A
practice transformer protection scheme the theoretical basis”, IEEE Trans. Electr. Insul., vol. 23,
transformer failure rate, estimated through risk no. 3, pp. 373-381, 1988
assessment, is reduced by about 11%. Surge [16] Z. G. Datsios, P. N. Mikropoulos and T. E. Tsovilis:
protective devices provided the necessary “Insulator string flashover modeling with the aid of
protection to the distribution transformer against an ATPDraw object”, in Proc. 46th UPEC, Soest,
transferred overvoltages. To further improve the Germany, 2011
lightning performance of the transformer additional [17] W. A. Chisholm, J. G. Anderson, A. Phillips and J.
measures should be considered such as reducing Chan: “Lightning performance of compact lines”, in
the transformer grounding resistance and installing Proc. X SIPDA, Curitiba, Brazil, 2009, pp. 45-64
line arresters; this calls for further work. [18] A. R. Hileman: “Insulation coordination for power
systems”, Boca Raton, Florida, CRC Press, 1999
6 REFERENCES [19] IEEE Guide for improving the lightning performance
of electric power overhead distribution lines, IEEE
[1] R. L. Witzke and T. J. Bliss: “Surge protection of Std 1410-2010, 2011
cable-connected equipment”, Trans. AIEE, vol. 69, [20] M. Kizilcay and C. Neumann: “Lightning overvoltage
no. 1, pp. 527-542, 1950 analysis for a 380-kV gas-insulated line”, in Proc.
[2] T. J. Carpenter, I. B. Johnson and L. E. Saline: IPST, Delft, The Netherlands, 2011
“Evaluation of lightning-arrester lead length and [21] Z. G. Datsios, P. N. Mikropoulos and T. E. Tsovilis:
separation in co-ordinated protection of apparatus “Impulse Resistance of Concentrated Tower
against lightning”, Trans. AIEE, vol. 69, no. 2, pp. Grounding Systems Simulated by an ATPDraw
933-944, 1950 Object”, in Proc. IPST, Delft, The Netherlands, 2011
[3] R. L. Witzke and T. J. Bliss: “Co-ordination of [22] P. N. Mikropoulos, T. E. Tsovilis, Z. Politis and A. G.
lightning arrester location with transformer insulation Kagiannas: “Evaluation of fast-front overvoltages
level”, Trans. AIEE, vol. 69, no. 2, pp. 964-975, 1950 arising at a 20/0.4kV distribution transformer”, in
[4] S. S. Kershaw, Jr. and C. R. Clinkenbeard: Proc. 7th MedPower, Agia Napa, Cyprus, 2010
“Discharge voltage of arrester connecting lead [23] D. P. Agoris, A. Stamatelos, E. C. Pyrgioti, D.
wires”, IEEE Trans. Power App. Syst., vol. PAS-93, Vasileiou and S. Dragoumis: “Post mounted
no.1, pp. 226-232, 1974 distribution transformer failures due to lightning
[5] G. L. Goedde, L. A. Kojovic and J. J. Woodworth: correlating to the grounding resistance”, in Proc.
“Surge arrester characteristics that provide reliable 28th ICLP, Kanazawa, Japan, 2006, paper no. VI-26
overvoltage protection in distribution and low- [24] IEC 60060-1, High-voltage test techniques-Part 1:
voltage systems”, in Proc. IEEE Power Eng. Soc. General definitions and test requirements, 2010
Summer Meeting, Seattle, USA, 2000, vol. 4, pp. [25] Lightning and Insulator Subcommittee of the T&D
2375-2380 Committee: “Parameters of lightning strokes: A
[6] F. D. Martzloff and K. Phipps: “Lingering lead length review”, IEEE Trans. Power Del., vol. 20, no. 1, pp.
legacies in surge-protective devices applications”, 346-358, 2005
IEEE Trans. Power Del., vol. 19, no. 1, pp. 151-157, [26] P. N. Mikropoulos and T. E. Tsovilis: “Statistical
2004 method for the evaluation of the lightning
[7] IEC 60071-2, Insulation co-ordination, Part 2: performance of overhead distribution lines”, IEEE
Application guide, 1996 Trans. Dielectr. Electr. Insul., vol. 20, no. 1, pp. 202-
[8] CIGRE-CIRED WG C4.402: “Protection of medium 211, 2013

308
View publication stats