Guidelines for Selection of Neutral Reactors rating

for Shunt Compensated EHV Transmission Lines

Veerabrahmam Bathini Nagaraja R K.Parthasarathy
Sr.Engineer-PSS Managing Director Technical Advisor
M/s PRDC Pvt.Ltd M/s PRDC Pvt.Ltd M/s PRDC Pvt.Ltd
Bangalore, India Bangalore, India Bangalore, India
veerabrahmam@prdcinfotech.com nagaraja@prdcinfotech.com

Abstract— Neutral reactors are generally employed in shunt
compensated long EHV transmission line to limit resonance
overvoltages induced on de-energized conductors due to parallel
energized circuits and stuck breaker conditions, and to reduce
the secondary arc current during single phase auto-reclosing.
The objective of this paper is to provide the guidelines for
selection of properly rated neutral reactors for shunt
compensated EHV transmission lines by conducting system
studies. These guidelines are demonstrated through a 360 km,
400 kV double circuit line with 80 MVAr shunt reactor at both
ends of each circuit. Studies were conducted using MiPower
power system analyses package.
Keywords-induced voltage; nuetral reactor; recovery voltage;
secondary arc current; single pole switchng; stuck breaker.
I. INTRODUCTION
S HUNT reactors are generally provided on long EHV
transmission lines to limit overvoltages during line
energization, load rejection and light load conditions. These
reactors are typically rated to compensate about 20 to 70% of
the line shunt capacitance. Although they limit overvoltages
under the above conditions, the shunt reactors could actually
increase voltages induced onto de-energized line conductors,
due to resonance from the energized conductors of same circuit
or another circuit on the same right of way. These overvoltages
could be limited by means of a reactor, termed a neutral
energized phases of the same circuit, but also by the mutual
coupling of the other healthy circuit.
The magnitude of the secondary arc current and the
recovery voltage are the most important factors, which
determine whether or not the secondary arc will be self-
extinguishing. Use of properly rated NGR at the neutral point
of the shunt reactor ensures secondary arc extinction and
successful SPS [3].
The parameters of NGR are initially determined based on
steady state analyses, considering equivalent networks on
either side of the transmission line under consideration.
Appropriate rating of the NGR is then selected by performing
transient analysis studies considering arc modeling, arc
extinction time and single line to ground fault at different
points on the transmission line. Both steady state and transient
analyses studies have to be performed using EMTP-type
program.
II. SELECTION OF NGR RATING
A. Equivalent network model for simulations
Equivalent network model is derived based on load flow
and short circuit study results and a two bus equivalent system
is formed as shown in Figures 1 and 2 (single circuit or double
circuit configuration).
SS-A SS-B

grounding reactor (NGR), connected between the shunt reactor C

B
C C
T
T
top
C A
T
top

T
A B
T
T
top
B C C
C

neutral and ground. The majority of transmission line faults are A

B B
T
middle
B C
middle
T
C A
middle
T
A B B
A
bottom bottom bottom

temporary short circuits. Automatic single phase reclosing is A A A B B C C A A

used to clear single-phase-to-ground faults, which are about

Rs

Rs

Rs

Rs

Rs

Rs
Ls

Ls

Ls

Ls

Ls

Ls
80% of the transient faults. The short circuit arc is usually self-
Rn

Rn

extinguishing after opening the transmission line circuit

+

+
Ln

Ln
+

breakers. High-speed re-closure of transmission line circuit

breakers can improve system stability. As the voltage level Figure 1. Equivalent network model for a typical single circuit
increases arc de-ionization time increases as well, endangering
system stability. Application of automatic single phase
+

+
Ln

Ln

reclosing makes it possible to increase system stability even for

+

+
Rn

Rn

extremely high voltage transmission lines. To enable successful

Ls

Ls

Ls

Ls

Ls

Ls

fast reclosing, NGR is normally used when transmission line is

Rs

Rs

Rs

Rs

Rs

Rs

SS-A SS-B

compensated with shunt reactors [3]. C

B
C C A
T
top

T
A B
T
top

T
B C
T
top

T
C C
C

B
B B B B C C A A B B
middle middle middle
A T T T A

During the dead time of Single Pole Switching (SPS), A A C

bottom
C A
bottom
A B
bottom
B A A

extinction of the main transient arc current should take place.

However, the faulted phase remains capacitively and C

B
C C C
T
top

T
C A
T
T
top
A B
T
top

T
B C C
C

inductively coupled with the energized un-faulted phases A

B B B
middle

T
bottom
B C
middle

T
bottom
C A
middle

T
bottom
A B B
A

A A A A B B C C A A
resulting in continuation of the fault current, which are known
Rs

Rs

Rs

Rs

Rs

Rs

as secondary arc currents. In case of SPS on double circuit

Ls

Ls

Ls

Ls

Ls

Ls

EHV line, the secondary arc is maintained not only by the

Rn

Rn
+

inter-phase capacitive coupling between faulted phases and

Ln

Ln
+

Figure 2. Equivalent network model for a typical double circuit

1
 Equivalent sources at sending and receiving end are
represented using positive and zero sequence impedance values
corresponding to the expected fault levels at both ends. The
single/double circuit transmission line is modeled by 3x3/6x6
phase impedance matrix (self and mutual impedances) and a
3x3/6x6 phase admittance matrix (self and mutual
capacitance). The investigation of NGR application is
performed for actual transposition (typical transposition of
single circuit and double circuit are shown in Figures 1 and 2
respectively). Transmission line shunt reactor is also modeled
as 3x3 phase impedance matrix (self and mutual impedances)
to consider mutual coupling among the phases of reactor if any.
Normally these reactors do not have mutual coupling.
B. Steady State Analysis
Various simulation studies have to be conducted to select
the initial value of NGR. Only steady state conditions have to
be considered in these simulations. The various studies to be
conducted are as follows.
1) Single pole switching
Single line to ground (SLG) faults account for 70%-95% of
faults on EHV transmission lines and most of these are
transitory in nature. From the stand point of minimizing the
disturbances, especially loss of synchronism which may
hamper the system stability caused by SLG fault, as well as to
maintain reliability, it is desirable to clear them by opening
only the circuit breaker pole on both terminals of the faulted
phase out of the three phases and re-close after a certain time
gap. This allows two energized and healthy phases to continue
carrying power during the period of interruption, which has
significant benefits. The increasing difficulty of construction of
new EHV transmission lines as well as high cost makes the
SPS an attractive method of achieving reliable power delivery
system [5].
The network to be considered for steady state analysis is
shown in Fig. 1 or 2 based on transmission line configuration.
SLG fault is created at different locations viz., SS-A end,
midpoint and SS-B end etc., with the NGR values varying from
0% to 100% of the phase reactor value. Single pole switching
namely viz. opening of the single phase of the breakers at both
the ends of circuit, is then carried out.
Values of the following parameters for different simulation
cases are to be recorded.
 Steady state primary arc current
 Steady state secondary arc current
 Steady state recovery voltage
 Rate of Rise of Recovery Voltage (RRRV) in
kV/ms after the SPS operation.
 Steady state neutral voltage and current for all the
line reactors
2) Induced voltage during stuck breaker conditions
Open phase conditions may occur when line single-phase
recloser is applied, or at the occurrence of stuck breaker poles
in the opening or closing operation. Series resonance can occur
in shunt compensated transmission lines during unbalanced
switching operations resulting in open phase overvoltages that
can damage line connected equipments and can therefore affect
system security and availability. In open-phase condition a
series resonance may occur with the coupling capacitance to
the energized phases and large overvoltages may stress the
open phases and associated open circuit breakers at their
terminals [3].
EHV breakers are usually designed to operate with single
pole mechanisms. It is possible that due to mechanical
differences or defects, all three poles may not operate
simultaneously or one of them can get stuck. One phase could
be left open with the other two phases energized during stuck
breaker condition while energizing the line or a single pole
open condition arises while performing single pole reclosing.
Similarly, two phases could be left open with the other phase
energized during line de-energization. Shunt reactors increase
the open-phase voltage considerably because of unequal
compensation of the positive and zero-sequence line
capacitances. As reactors are in parallel with the line conductor
capacitance to ground, the equivalent phase-to-ground
reactance at power frequency is inductive and may reach high
value when the shunt compensation is large (above 65%). In
such cases, parallel combination of the shunt reactor and the
line shunt capacitance in series with the inter-phase capacitance
forms a series resonant circuit. These conditions could result in
series resonance on shunt compensated lines with attendant
overvoltages and their detrimental effects on the connected
equipments [6].
Steady state analysis of equivalent network model has to be
performed for various simulated stuck breaker conditions.
Details of one stuck breaker condition to be simulated are
as follows.
 Energize transmission line from SS-A to SS-B with
one pole of SS-A end breaker with one pole stuck and
keeping open all the three poles of breaker at SS-B
end.
 Vary NGR value from 0% to 100% of phase reactor
value.
 Record steady state open conductor(s) phase to
ground voltage.
 Record steady state neutral voltage and current for all
line shunt reactors.
 Observe whether there is any possibility of getting
resonance at specific value of NGR.
Perform other stuck breaker simulated case studies.
3) Induced voltages on de-energized circuit
This study is applicable for only double circuit lines. A
shunt compensated de-energized circuit running on the same
right of way with an energized circuit can be subjected to high-
induced voltages due to parallel resonance between the line
shunt reactor and the line capacitance. The phenomenon of
induced voltages, due to electrostatic and electromagnetic
coupling, on a shunt compensated de-energized circuit from a
parallel-energized circuit needs to be studied.
As the NGR is connected at the neutral point of the phase
reactor, this may lead to a sustained oscillation in the ring
down voltage on the de-energized circuit in a double circuit
line. In the extreme case, depending upon the line length and
degree of shunt reactive compensation on the line, resonating
overvoltages may occur. Keeping this aspect in view the NGR
value selected has to be examined for possible occurrence of
resonating overvoltages.
Equivalent system to be considered for this induced voltage
study is shown Fig. 2. Open all the three poles of circuit
2
breakers at both ends of one circuit and keep the other circuit in
energized condition. Vary NGR value from 0% to 100% of
phase reactor value.
For each NGR value the following values have to be
recorded.
 Steady state phase to ground voltages of de-
energized circuit.
 Steady state neutral voltage and current for all line
shunt reactors.
 Observe whether there is any possibility of
resonance for this value of NGR.
C. Selection of initial NGR value:
Based on the steady state analyses performed, the initial
value of NGR is selected keeping in view the following two
desired check conditions.
 Successful secondary arc extinction is normally
expected for single/double circuit EHV line, if the
secondary arc current is less than 40 A and the rate of
rise of recovery voltage (RRRV) is less than 10
kV/msec [4]. The NGR value selected should satisfy
these criteria.
 Based on the stuck breaker condition and induced
voltages on de-energized circuit studies, examine
whether there is any possibility of occurrence of
resonance for the NGR value selected.
After checking that the NGR value selected satisfies both
check conditions, transient analysis studies namely,
energization, load rejection, single pole switching (open and
reclose), and induced voltage studies are conducted. The rating
of NGR is finalized based on the results of these studies.
D. Transient Analyses
1) Single Pole re-closure study
Equivalent system to be considered for the simulation is as
per Fig. 1 or 2. Perform transient analysis study by creating
SLG fault at various locations on the transmission line. Arc has
to be modeled by variable fault conductance so that arc
extinction is automatically controlled. Select suitable value for
dead time to enable successful single pole re-closure [1], [2].
For each fault location the following waveforms have to be
recorded.
 Secondary arc current
 Recovery voltage
 Neutral voltage and current for all the shunt
reactors.
Observe the secondary arc extinction time and compare this
with specified dead time to ensure successful re-closure.
2) Temporary Overvoltage study (Load rejection and fault
clearing)
Other important parameters that are required to be specified
for NGR design are continuous current rating, short time
current rating, insulation level, etc. For determining the short
time current rating and insulation level, transient studies are
performed by simulating only load rejection, only SLG fault
and load rejection accompanied with SLG fault.
For each case the following waveforms have to be
recorded.
 Phase to ground voltages at SS-A and SS-B ends of
transmission line
 Neutral voltage and current for all the shunt
reactors.
Based on this study, select the insulation class, short time
current rating and maximum peak current of the NGR.
3) Switching Overvoltage Study
Conduct line energization studies with selected NGR value
to ensure that switching overvoltages are within acceptable
limits considering trapped charges and line surge arresters.
Check the energy class of surge arresters used.
4) Induced voltages on de-energized circuit with selected
NGR value (applicable for double circuit line)
Perform the induced voltage study on second circuit (de-
energized) by energizing the first circuit. Also conduct induced
voltage studies on de-energized circuit by creating SLG fault
on the energized circuit. Vary the MVAr value of the line
reactor in +/-10% range. For example, for 80 MVAr line
reactor vary the MVAr value of the line reactor from 72 to 88
MVAr in suitable steps. Also consider +/- 2.5% manufacturing
tolerance for line reactor values amongst the phases.
For each case the following waveforms are to be recorded.
 Phase to ground voltages on de-energized circuit.
 Neutral voltage and current for all the line reactors.
Based on this study, determine the continuous current
rating for NGR value selected.
E. Summary of Studies
Observations made from Summary of studies are presented
in Table I.
TABLE I. SUMMARY OF SYSTEM STUDIES
Summary of studies
Parameter Selection Criteria
Rated Impedance Based on steady state analyses (SPS, stuck
breaker and induced voltages studies) NGR
value „x‟ % of Xs is selected. Considered a
manufacturing tolerance of +/-2.5%. Select the
NGR rated impedance to be nearest to the
standard value available from manufacturer.
Rated For 10 sec Based on steady state analyses and load rejection
current studies observe the maximum neutral current
and and arrive at the 10 seconds short time current
voltage and voltage rating for NGR.
Continuous Considering manufacturing tolerance of +/-2.5%
current and for main reactor and system voltage unbalance
voltage of 1.5% for EHV system, calculate the
maximum continuous current flowing through
the neutral of line reactors. Check this value
with the maximum neutral current value
recorded while performing Induced voltages on
de-energized circuit transient analysis. Also
check this value with 3% of 10 second current
rating (IEEE std., 32-1972) computed earlier.
Select the continuous current and voltage rating
for the NGR based on whichever value is
maximum.
3
 Rated peak current Based on Transient analysis studies observe the Recovery Voltage at Midpoint
maximum initial asymmetric peak current RRRV at midpoint

Rate of rise of recovery

200 20

Recovery Voltage
flowing through neutral.

voltage (kV/ms)
160 16

(kVrms)
Calculate the asymmetric peak current based on 120 12
IEEE Std 32-1972. 80 8
Comparing these two values select the 40 4
maximum to be the rated peak current for NGR. 0 0
Voltage class and Based on the steady state and transient analysis 0% 10% 20% 30% 40% 50% 60%
Insulation level at neutral studies observe the maximum neutral voltage. NGR (Xn/Xs) value
point. By using this value and referring to IEEE std,
32-1972 select the insulation class for NGR Figure 4. Recovery voltage and rate of rise of recovery voltage for a fault
based fault voltage criteria. considered at the midpoint

Surge arrester rating for Duty cycle of surge arrester is chosen based on 60
NGR temporary overvoltage study or 10 second

Current (Arms)
Secondary Arc
50
voltage rating. 40
30
Various parameters to be recommended for NGR rating are 20
presented in Table III for a typical case study. 10
0
III. CASE STUDY 0% 10% 20% 30% 40% 50% 60%
A. System Data NGR (Xn/Xs) value

Figure 5. Secondary arc current for a fault considered at the midpoint

1) Equivalent network data
The equivalent source impedance data considered for study From Figures 4 and 5, 30% NGR value gives recovery
are presented in Table II. voltage of 40 kVrms, RRRV of 6.5 kV/ms and the secondary
TABLE II. EQUIVALENT SOURCE IMPEDANCE DATA
arc current of 20 Arms.

Equivalent Source Impedances 2) Stuck breaker condition

SS-A SS-B Different stuck breaker condition has been considered and
R1 () 0.4 4.0
X1 () 8.0 80.0
simulated to arrive the initial NGR rating. Results of a typical
R0 () 2.0 20.0 stuck breaker condition as shown in Fig. 6 are presented in
X0 () 24.0 160.0 Figures 7 and 8.
2) Transmission line data
The 360 km, 400kV double circuit transmission line data
considered for the study is presented in Fig. 3. Transposition
for the transmission line is considered based on Fig. 2 and a
shunt compensation of 80 MVAr has been considered at both
ends of each circuit for the studies.

Figure 6. Equivalent model for a typical stuck breaker condition

Phase voltage at midpoint

Neutral voltage at SS-A end
1000 100
Open phase voltage

Neutral Voltage

800 80
(kVrms)
(kVrms)

600 60
400 40
200 20
0 0
Figure 3. Transmission line data 0% 10% 20% 30% 40% 50%
NGR (Xn/Xs) value
3) Surge arrester data Figure 7. Open phase voltage and neutral voltage for a typical stuck breaker
condition
The 360 kV, class 4 surge arrester V-I characteristics are
referred from [7].
600
Neutral Current

B. Steady Stae Analysis 500

400
(Arms)

1) Single pole switcing 300

200
Single pole switching study results for a typical case of a 100
fault considered at the midpoint of the transmission line 0
between substation A and B are presented in Figures 4 and 5. 0% 10% 20% 30% 40% 50%
Similarly, studies need to be conducted for faults at different NGR (Xn/Xs)
locations to arrive at the initial NGR rating. Figure 8. Neutral current for a typical stuck breaker condition
4
 From Figures 7 and 8, 30% NGR value gives open phase
voltage of 100 kVrms, neutral voltage of 76 kVrms and neutral
current of 127 Arms.
3) Induced voltage
Studies regarding induced voltages on de-energized circuit
have been simulated for different cases with the other circuit
energized and with considering fault on the energized circuit at
different locations. The simulation result for one typical case is
presented in Fig. 9.
Induced voltage on de-energized circuit
(b)
With fault at SS-A end on energized circuit
Induced Voltage with fault on

Other circuit is energized

energized circuit (kVrms)

Induced Voltage with other

circuit enegized (kVrms)
500 30
400 28
300 26
200 24
100 22
0 20
0 10 20 30 40 50 60 70 80 90
% NGR (Xn/Xs) value

Figure 9. Induced voltages on de-energized circuit

(c)
From Fig.9, 30% NGR value gives induced voltage on de- Figure 10. (a)Secondary arc current , (b) Neutral current and (c) Neutral
energized circuit 23 kVrms and 150 kVrms during normal and voltage for a fault considered at SS-B end
fault conditions respectively.
From Fig. 10 (a), 30% NGR value gives the dead time of
C. Section of Intial NGR Value
250 ms, Fig.10 (b) shows the neutral current peak of 318 A.
Based on literature [4], successful secondary arc extinction Fig.10(c) shows the maximum neutral voltage of 65 kVrms,
would apparently be expected for double circuit EHV line, if based on IEEE Std.32 using fault voltage criteria 69 kVrms
the secondary arc current is less than 40 A and the rate of rise insulation class selected at neutral point.
of recovery voltage (RRRV) is less than 10kV/msec. Based on
this reference and considering safety margin it is recommended 2) Temporary Overvoltage study (load rejection)
to select NGR value as 30% of Xs for successful secondary arc Simple load rejection and load rejection accompanied with
extinction. Also, from stuck breaker studies and studies single line to ground fault were conducted on both ends of
regarding voltages induced on de-energized circuit, it is substations separately. Results of these studies are used for
observed that the induced voltages are within acceptable limits selection of surge arrester rating for protection of neutral
for NGR value of 30%. Hence, initially an NGR value of 30% reactor.
of Xs is selected and is considered for further studies namely,
load rejection, switching, transient analysis-single pole 3) Swicthing Overvoltages
reclosing, and induced voltages based on which NGR rating is
finalized. Switching overvoltage studies have been conducted for
different operating conditions and the results for the same are
D. Transient Analysis presented in Fig. 11.
1) Single pole switching
Line Voltage Profiles during line energization
Single pole switching study results for a typical case of a
fault considered at the substation B are presented in Fig. 10. without SA and NGR only with SA
Similarly, studies need to be conducted for faults at different only with NGR with SA and NGR
locations to finalize the NGR rating.
3
2.5
Voltage (p.u)

2
1.5
1
0.5
0
0 120 180 240 360

Distance from SS-B end (km)

(a) Figure 11. Line voltage profile during line energization

5
 From Fig.11, it is observed that with 30% NGR and 360 kV TABLE III. RECOMMENDED NGR PARAMETERS
surge arrester switching overvoltages are within limits. Recommended design rating for NGR
Rated Impedance 600 ohms
4) Induced voltage
Rated Current 140Arms, for 10 seconds
Studies regarding induced voltages on de-energized circuit
10 Arms, Continuous
have been simulated for different cases by varying the MVAr Rated Voltage 84.0 kVrms, for 10 seconds
value of the line reactor in +/-10% range and +/- 2.5% 6.0 kVrms, for continuous
manufacturing tolerance for line reactor values amongst the Rated peak current 510 Apeak
phases at different locations. The simulation results for the
same are presented in Fig. 12. Rated Power 11760 kVAr for 10 seconds
60 kVAr Continuous
Rated Frequency 50 Hz
Induced phase voltage on de-energized circuit
With other circuit energized
Rated Insulation Class at neutral point 69 kVrms
Phase Voltage (kVrms)

With fault on energized circuit Minimum BIL value at neutral point of 350 kVpeak
120 reactor
100 Minimum BIL value at neutral bushing of 380 kVpeak
80 reactor
60 No. of Phases 1
40 Connection Single Phase/Neutral
20
Insulation class at earthing side 15 kVrms
0
BIL Earthing Side 110 kVpeak
72 76 80 84 88
Shunt reactor MVAR value Temperature Rise To be specified by Vendor
Surge Arrester, class 78 kVrms rated voltage, class 4
(a)

Neutral voltage on de-energized circuit

With fault on energized circuit IV. CONCLUSIONS
With other circuit energized Shunt reactors are generally provided on long EHV
Voltage (with other circuit

50 1.2
energized circuit) kVrms
Voltage (with fault on

transmission lines to limit overvoltage during line energization

energized) kVrms

40 1
and load rejection. Use of neutral grounding reactor at the
30
0.8 neutral point of the shunt reactor on long EHV lines is
20
judiciously adopted to ensure the secondary arc extinction and
10 0.6
successful single pole switching as 80% of the transient faults
0 0.4 on EHV lines are single line to ground fault. This paper
72 76 80 84 88 provides the guidelines to properly rate the NGR for the steady
Shunt reactor MVAR value
state and transient duties to which they will be exposed by
(b) determining the NGR parameters initially based on steady
Neutral current on de-energized circuit
state analysis and finalizing the ratings by performing transient
With fault on energized circuit analysis with arc modeling and arc extinction time.
With other circuit energized
90 2.5
REFERENCES
energized circuit) Arms

Current (with other circuit

Current (with fault on

80 2
energized) Arms

70
60
50
72 76
Shunt reactor MVAR value
Figure 12. (a)Induced phase voltage (b) Neutral voltage at shunt reactor
neutral (c) neutral current through shunt reactro on de-energized circuit.
From Fig.12, for selected 30% NGR value, induced
voltages on de-energized circuit are within the limits of rated
voltages. Neutral current of 2 Arms for normal condition used
as one of the parameter to select the continuous rating for
NGR.
E. Recommended NGR Parameters
Based on steady state and transient analyses studies the
recommended NGR parameters are presented in Table III.
1.5 [1] A.T. Johns, R.K. Aggarwal, and Y.H. Song, “ Improved techniques for
modeling fault arcs on faulted EHV transmission systems”, IEEE Proc.
1 Generation, Transmission, and Distribution, Volume 141, Issue 2, March
1994, pp: 148-154.
0.5
[2] A.I. Megahed, H.M.Jabr, F.M. Abouelenin, and M.A.Elbakery, “Arc
80 84 88 characteristics and a single-pole auto-reclosure scheme for alexandria
HV transmission system”, International Conference on Power system
Transients-IPST 2003 in New Orleans, USA.
(c) [3] S.R Atmuri, R.S Thallam "nuetral reactors on shunt compensated ehv
lines" Transmission and Distribution Conference, 1994., Proceedings of
the 1994 IEEE Power Engineering Society.
[4] Gary C. Thomann, SM , Stephen R. Lambert, F , Somkiet Phaloprakarn ,
“non-optimum compensation schemes for single pole reclosing on EHV
double circuit transmission lines”, IEEE Transactions on Power
Delivery, Vol. 8, No. 2, April 1993, pp:651-659.
[5] Mohammad Reza Dadash Zadeh , Majid Sanaye-Pasand, “ Investigation
of Neutral Reactor Performance in Reducing Secondary Arc Current”,
IEEE Transactions on Power Delivery, 2008.
[6] F. ILICETO, E. CINIERI, A.DIVITA,” Overvoltages due to open-phase
occurrence in reactor compensated EHV Lines”, IEEE Transactions on
Power Apparatus and Systems, Vol.PAS-103, No.3, March 1984.
[7] ABB SURGE ARRESTERS — BUYER‟S GUIDE, Edition 6, 2008-08.