Interaction of the secondary arc

Interaction of the Secondary Arc with the Transmission
System during Single-Phase Autoreclosure
M. Kizilcay, Member, IEEE, G. Bán, Fellow Member, IEEE, L. Prikler, Member, IEEE,
and P. Handl, non-member
Abstract -- Secondary arcs can be modeled in a numerical
transients program like ATP-EMTP. The dynamic behavior of
the arc is represented as a time-varying resistance by solving a
differential equation of arc conductance based on the energy
balance in the arc column. The almost random variation of arc
parameters influences significantly the arc performance during
single-phase autoreclosure on transmission lines. Depending on
the variation of arc parameters, certain effects of interaction
with the transmission system are observed. In this relation, field
tests on real overhead lines are extremely useful to verify
numerical arc simulations and to understand the arcing
phenomenon. Authors have participated in the past in several
field tests performed on a 400 kV overhead line without shunt
compensation and on 750 kV line with shunt compensation.
Index Terms – autoreclosure, arc, arcing fault, EMTP,
primary arc, secondary arc, shunt reactor, transients,
transmission line.
I. INTRODUCTION
AULTS on EHV/UHV overhead lines are in majority
single-phase-to-ground arcing faults and are temporary in
most cases. Single-phase auto-reclosure (SPAR) is an
effective measure to eliminate arcing faults. The fault arc is
classified according to the fault state regarding the fault
current. Primary arc is effective after fault inception till
single-phase tripping of the faulty phase. Secondary arc
follows the primary arc after isolating the fault by single-
phase tripping on a transmission line. It is sustained by the
capacitive and inductive coupling to the sound phases. The
secondary arc self-extinguishes usually, but its life-time may
have a strong influence to the operation of the line: the longer
the secondary arcing time, the higher the risk of unsuccessful
reclosing. Lengthening the SPAR dead time is limited by
dynamic stability constraints. This limit is usually less than
1.5 - 2 seconds for long EHV/UHV interconnections.
Whereas the primary arc shows generally a deterministic
behavior as observed at field and laboratory arc tests [1], [2],
the secondary arc has extremely random characteristics
effected by the external conditions around the arc channel like
M. Kizilcay is with the Dept. of Elect. Eng. & Comp. Sci., Osnabrueck
University of Applied Sciences, Germany (e-mail: kizilcay@fhos.de)
G. Bán is with the Dept of Electric Power Engineering, Budapest
University of Technology and Economics (e-mail: ban@vmt.bme.hu)
L. Prikler is with the Dept of Electric Power Engineering, Budapest
University of Technology and Economics (e-mail: l.prikler@ieee.org)
P. Handl is PhD student at the Budapest University of Technology and
Economics (e-mail: handlp@vmt.bme.hu)
ionized surrounding air, wind, thermal buoyancy and electro-
dynamic force. In spite of this fact, a numerical arc model may
be an useful tool to identify the main influencing factors and
the interaction of the arc with the electric circuit, and to guess
worst-case arcing time.
II. MODEL OF THE FAULT ARC
The arc model used in this work is based on the energy
balance of the arc column and describes an arc in air by a
differential equation of the arc conductance g [2], [3]:
( )
1dg
G g
dt τ
= − (1)
where τ: is the arc time constant,
g: instantaneous arc conductance,
G: stationary arc conductance.
The stationary arc conductance is defined as:
arc
st
i
G
u
= (2)
with
( )0 0' ' ( )st arc arcu u r i l t= + ⋅ (3)
where iarc: instantaneous arc current,
ust: stationary arc voltage,
larc: instantaneous arc length,
u'0: characteristic arc voltage per length,
r'0: characteristic arc resistance per length.
Equation (1) is a generalized arc equation that is suitable
to represent an arc in an electric circuit by a two-pole. The
parameters τ, u'0 and r'0 can be obtained from arc
measurements. In case of secondary arc they vary significantly
with arc length larc. The dependence of the arc time constant τ
on larc can be defined by the relation
0
0
arcl
l
α
τ τ
 
= ⋅ 
 
(4)
where τ0: initial time constant,
l0: initial arc length,
α: coefficient of negative value.
F
0-7803-7967-5/03/$17.00 ©2003 IEEE
Paper accepted for presentation at 2003 IEEE Bologna Power Tech Conference, June 23th-26th, Bologna, Italy

Since the length variation of the secondary arc is highly
dependent on external factors like wind, thermal buoyancy, it
is difficult to consider these random effects accurately in the
arc model. Nevertheless the above arc model may be used
successfully to estimate the maximum arcing time as worst
case or to understand the interaction of the secondary arc with
the electrical circuit.
A. Arc Representation in EMTP-ATP
The arc as a nonlinear dynamic element is represented by
the Thevenin type, Type-94 component in the ElectroMagnetic
Transients Program EMTP-ATP [4]. The arc is described in
MODELS language. In the Thevenin type component the
electric arc as a two-pole sees the rest of the linear circuit as a
Thevenin equivalent. Inputs are Thevenin voltage vth and
resistance rth at the terminals. The arc model calculates the
value of the resulting arc current iarc at each time step. The
interaction of the electric arc with the remaining circuit is
shown in Fig. 1.
rth
gvth
Electric circuit
iarc iarc
vth
rth
MODELS
arc description
Fig. 1. Interaction between the electic circuit and the arc model
At each time step first the stationary arc voltage ust and
time constant τ are updated using (3) and (4) which depend on
instantaneous arc length larc. The arc current iarc can be
expressed referring to Fig. 1 as follows
1
th
arc
th
g v
i
g r
⋅
=
+ ⋅
(5)
The arc equation (1) is solved using MODELS's
LAPLACE function to obtain g:
)(
1
1
)( tG
s
tg ⋅
⋅+
=
τ
(6)
Equations (2), (5) and (6) are solved simultaneously using
an iterative method available in MODELS as "COMBINE
ITERATE … ENDCOMBINE" [4]. The use of type-94 component
in EMTP-ATP enables simultaneous solution of arc equations
together with the equivalent system of the electric network.
B. Criteria for Self-Extinction of the Arc
The secondary arc elongation causes increase of arc
voltage. If the arc elongation is rather smooth without steep
rise, for example due to gust, then the secondary arc
extinguishes as the arc voltage reaches the level of recovery
voltage [5]. The recovery voltage is the voltage across the
fault path at arc location after extinction of the secondary arc.
Consequently, the arc duration may be estimated by the
elongation speed of the arc.
Besides the arc elongation, re-ignitions inside the arc
channel may have a significant role in the process. The speedy
elongation of the arc separates the conducting plasma clouds
by high resistance channel zones. So the resistance of the
whole arc length increases. It often happens however, that the
recovery voltage makes the high resistance arc channel zones
re-ignite during the arc elongation process. If the recovery
voltage is sufficient to produce a breakdown, bridging a
significant length of the arc channel, the arcing process may
return to the steady-state condition, remarkably prolonging the
self-extinction time.
The arc self-extinction phenomenon is not known in detail.
There are two different approaches that try to explain arc
extinction: A) based on thermal instability described by arc
equation (1). The arc extinguishes, if the time derivative of
instantaneous arc resistance, drarc/dt, exceeds a pre-defined
limit provided arc conductance g' per length is less than g'min.
B) based on dielectric phenomenon [6]. It is assumed that
following each arc current reversal, the arc would first
extinguish. The arc current is held at zero for as long as the
arc recovery voltage remains below an arc re-ignition voltage
characteristic obtained empirically.
C. Random Variation of Arc Parameters
Since the arc parameters are expressed as functions of the
instantaneous arc length, the random arc behavior can be
reproduced by varying the arc length in a random way. It is
physically reasonable to describe a global arc length increase
– either piecewise-linear or any predefined function –
superposed by a local random length variations that should
imitate the local breakdowns along the elongated secondary
arc. Fig. 2 shows an example of arc length variation that is
used to simulate fault arc whose voltage and current
waveforms were recorded at a field test performed on a 230
km long, 400-kV overhead line without shunt compensation
as shown in Fig. 3 [2]. The line has been operated by
connecting the two circuits in parallel along the 1/3rd
of the
full length. The remaining sections of the 2nd
circuit have been
disconnected from the tested line or were grounded.
Fig. 4 and 5 show the measured and computed arc voltages
and currents, respectively. The steady-state recovery voltage is
shown additionally in Fig. 4. When the secondary arc voltage
reaches the level of recovery voltage, the arc extinguishes. For
the arc simulation following arc data are used:
u'0 = 0.9 kV/m, τ0 = 1 ms
r'0 = 22 mΩ/m ; α = −0.5.
Initial length of the primary arc is assumed as 4 m. For arc
extinction the following limiting values per arc length were
used:
min' 50µS m,
'
20M /(s m)arc
g
dr
dt
= ⋅
= Ω ⋅

0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2[s]
0
3
6
9
12
15
[pu]
time
primary arc secondary arc
Fig. 2. Time-varying arc length with random local variations
90 km
104 km
phase to
ground arc 230 km
topen
= 0.568 s
(phase a)
Fig. 3. 400-kV transmission system with arc tests
0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2[s]
-100
-50
0
50
100
[kV] arc voltage recovery voltage
time
measurement
0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2[s]
-100
-50
0
50
100
[kV] arc voltage
computation
time
Fig. 4. Measured and computed arc voltages
0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2[s]
-500
-250
0
250
500
[A] primary arc current
secondary arc current
measurement
time
time0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2[s]
-500
-250
0
250
500
[A] primary arc current
secondary arc current
computation
Fig. 5. Measured and computed arc currents

Fig. 6. Sustained 400-kV secondary arc at calmness
III. EXCITATION OF TRAVELING WAVES ON THE LINE
BY THE SECONDARY ARC
In the field tests of secondary arc performance on a double-
circuit, 400-kV overhead line of 230 km length without shunt
compensation the authors from Budapest identified from
records of arcs (see Fig. 6) at calmness with long self-
extinction times that the arc re-ignitions (transition from a
very low arc conductance state to a high conductance state)
initiate traveling waves on the disconnected phase of the line.
The reflected waves at the far end of the line influence the arc
current shape and its pulse width.
0.83 0.85 0.87 0.89 0.91 0.93
-140
-105
-70
-35
0
35
70
-300
-150
0
150
300
450
600
[kV] [A]arc voltage
arc current
time [s]
Fig. 8. Measured secondary arc voltage and current
0.1267 0.1467 0.1667 0.1867 0.2067 0.2267[s]
-140
-105
-70
-35
0
35
70
[kV]
-300
-150
0
150
300
450
600
[A]arc voltage
arc current
time
Fig. 9. Computed secondary arc voltage and current
The EMTP-ATP model of the test arrangement has been
created using ATPDRAW graphical preprocessor [7] and is
shown in Fig. 7. Fig. 8 and 9 shows measured and computed
secondary arc voltages and currents in a time interval where
the arc is elongated suddenly and consequently the arc voltage
and time constant changes rapidly, respectively.
In the simulation the arc time constant is represented by
the data (u'0 and r'0 remain unchanged) :
τ0 = 0.8 ms, α = −1.6.
Fig. 7. ATPDRAW circuit of the 400-kV system including the arc model

0.1267 0.1467 0.1667 0.1867 0.2067 0.2267[s]
0.0
0.2
0.4
0.6
0.8
0.0
0.5
1.0
1.5
2.0
2.5
3.0
time
arc time constant
arc length variation
[ms]
[pu]
Fig. 10. Arc length and arc time constant variation in the simulation shown in
Fig. 9
0.1267 0.1467 0.1667 0.1867 0.2067 0.2267[s]
0
2
4
6
8
10
12
time
[mS]
Fig. 11. Impulse-like varying arc conductance
0.875 0.879 0.883 0.887 0.891 0.895[s]
-300
-200
-100
0
100
200
300
-110
-100
-90
-80
-70
-60
-50
[kV][A]
recovery voltage with
injected current pulses
Fig. 12. Correlation between traveling waves and shape of the arc current
These current pulses initiate traveling waves on the line.
The superposition of reflected waves produces a sudden
current zero resulting in a increase of arc column resistance.
In order to show the correlation between traveling waves
and the shape of arc current, an additional simulation has been
performed by injecting a current pulse with a very small width
(20 µs) into the line at the arc location by replacing the arc.
Reflections take place at 1/3rd of the line and at open ends of
the faulty phase. The simulation results with current pulse
injection superposed on the recovery voltage and a cycle of
measured arc current are shown in Fig. 12. The travel times
obtained by the EMTP-ATP model of the line match
practically with the changes in the waveform of the recorded
arc current.
IV. SECONDARY ARC INTERACTION WITH A SHUNT
COMPENSATED 750-KV LINE
The influence of the line-side connected shunt reactors on
the SPAR performance of the line was discussed in several
papers in the past [8], [9], [10]. The secondary arc model
presented in this paper is capable to reproduce the dynamic
interaction of the arc with the shunt compensated line.
Ls=3 H
Ln=0.3 H
Ls=6 H
0 0
1 2 3
C'pos=13.3 nF/km, C'0=9.7 nF/km
479 km
a
b
c
sending end receiving end
line configuration
Fig. 13. An arc fault on a shunt compensated 750-kV line
A single-line-to-ground arc fault is simulated at sending
end of a 479-km long, 750-kV single-circuit overhead line
where four-legged shunt reactor (with a neutral coil) is
installed on the line-side. The simplified diagram of the
transmission line is shown in Fig. 13. The phase conductors
consist of 4-bundle wires. The shunt reactor at the far end do
not possess any neutral coil. Following data are used for the
arc representation:
u'0 = 0.9 kV/m, τ0 = 0.8 ms
r'0 = 40 mΩ/m ; α = −0.5.
The initial arc length is assumed to be 15 m. A constant
elongation speed for the arc is chosen such that the arc length
will quadruple in 13 periods (0.26 s). The primary arc is
initiated at t = 0.01 s. The pole of phase a of circuit-breakers
at both line ends opens at t = 0.075 s. Fig. 14 shows the
computed arc voltage and current waveforms. The shunt
reactors produces a DC component in the initial part of the
secondary arc current producing asymmetry in the current
waveform of Fig. 14 around t = 0.2 s . The current of the
corresponding leg of the shunt reactor at the sending end of
the line is compared with the arc current in Fig. 15.
0.0 0.2 0.4 0.6 0.8 1.0 1.2[s]
-2.10
-1.75
-1.40
-1.05
-0.70
-0.35
0.00
0.35
0.70
[MV]
-300
-150
0
150
300
450
600
750
900
[A]
arc current
arc voltage
time
Fig. 14. Secondary arc voltage and current on the 750-kV line with four-
legged shunt reactor
0.00 0.05 0.10 0.15 0.20 0.25 0.30[s]
-200
-100
0
100
200
[A]
arc currentreactor current
time
Fig. 15. Arc current and current of the shunt reactor (phase a) at the sending
end of line

0.0 0.2 0.4 0.6 0.8 1.0 1.2[s]
-2.10
-1.75
-1.40
-1.05
-0.70
-0.35
0.00
0.35
0.70
[MV]
-300
-150
0
150
300
450
600
750
900
[A]
recovery voltage
arc current
arc voltage
time
Fig. 16. Secondary arc voltage, arc current and recovery voltage on the shunt
compensated line without neutral coil
Fourier analysis of the shunt reactor current (phase a)
shown in Fig. 15 in the interval (0.25 s … 0.27 s) results in
IDC = 6.5 A; I50Hz = 69.5 A (peak)
After arc extinction at t = 0.75 s the recovery voltage in
Fig. 14 shows low-frequency oscillations caused by the
interaction of the shunt reactors with line capacitances.
It is known that the neutral coil aims to shorten the duration
of the secondary arc by reducing the arc current and steepness
of the recovery voltage at the fault location [8]. Shunt reactors
without neutral coil has a negative impact on arc duration due
to the relatively high oscillating recovery voltage. For
comparison purpose a secondary arc simulation is performed
using the same 750-kV transmission system, except the
neutral coil at the sending end of the line is not taken into
consideration. The computed arc voltage and current are
shown in Fig. 16.
Additionally the recovery voltage is shown in this figure
that is obtained by a separate computation, where the arc is
replaced by a switch that opened at t = 0.1 s. Compared to the
case with neutral coil (Fig. 15) the secondary arc current and
recovery voltage are considerably higher causing a long
secondary arc duration. With the same arc model following
arc duration times are observed for both simulation cases:
- Shunt reactor with neutral coil : tarc = 0.67 s (100 %)
- Shunt reactor without neutral coil : tarc = 1.0 s (149 %)
The arc duration is expected to be 49 % longer, if the shunt
reactor at sending end is not equipped with a neutral coil. The
high amplitudes of the oscillating recovery voltage (Fig. 16)
may prolong the arc duration due to likely re-ignition of the
arc.
V. CONCLUSION
The arc model presented in this paper is capable to
replicate the interaction of an arc fault through air with the
remaining electrical network. Both arcing stages – primary
and secondary arc – can be represented during single-phase
autoreclosure as illustrated by simulation cases for an
uncompensated 400-kV line and shunt compensated 750-kV
overhead line.
The arc as a two-pole is modeled using type-94 component
in the EMTP-ATP that enables simultaneous solution of the
arc differential equation and electric circuit equations.
MODELS, a general-purpose simulation language integrated
in the EMTP-ATP, is used to describe the arc dynamics.
Due to highly random behavior of the secondary arc it is
difficult to reproduce exact arc duration by digital simulations.
In spite of this difficulty the arc model can be successfully
utilized to find main factors influencing the secondary arcing
process as illustrated in this paper by different simulation
cases. The arc interaction with the electric circuit that the
secondary arc may initiate traveling waves on the faulty phase
due to impulse-like shape of the arc current can be shown also
by digital simulations which are in agreement with the arc
measurements. The interaction of the secondary arc with the
line-side shunt reactors is shown for an existing 750-kV
transmission line. The influence of a neutral coil on the
secondary arc duration is illustrated by EMTP simulations.
VI. ACKNOWLEDGEMENT
The activity reported in this paper has been partly
supported by the Hungarian Research Fund under contract
OTKA T-035178. Authors acknowledge the contribution of
the Hungarian Power Companies Ltd. to the field
measurements.
VII. REFERENCES
[1] M. Kizilcay, K.-H. Koch, "Numerical fault arc simulation based on
power arc tests", ETEP Journal, vol. 4, no. 3, pp. 177-186, May/June
1994.
[2] L. Prikler, M. Kizilcay, G. Bán, P. Handl, "Improved secondary arc
models based on identification of arc parameters from staged fault test
records", presented at the 14th
Power System Computation Conf.,
Sevilla, Spain, June 2002.
[3] M. Kizilcay, T. Pniok, "Digital Simulation of Fault Arcs in Power
Systems", ETEP Journal, vol., 1, no. 1, pp. 55-60, 1991.
[4] Alternative Transient Program Rule Book, Can/Am EMTP User Group,
USA, 1997.
[5] H. J. Haubrich, et al., “Single-Phase Auto-Reclosing in EHV Systems”
CIGRE 1974, Rep. 31-09.
[6] A.T. Johns, R.K. Aggarwal, Y.H. Song, "Improved Techniques for
Modeling Fault Arcs on Faulted EHV Transmission System", Proc. IEE
– Generation, Transmission and Distribution, vol. 141, no. 2, pp. 148-
154, 1994.
[7] Prikler L, Høidalen HK. ATPDraw version 3.5 for
Windows9x/NT/2000/XP-User’s Manual. SINTEF Energy Research AS,
Norway, TR F5680, ISBN 82-594-2344-8, Aug 2002..
http://www.eeug.org/files/secret/atpdraw
[8] G. Bán, L. Prikler, G. Banfai, "The Use of Neutral Reactors for
Improving the Successfulness of 3-phase Reclosing", presented at the
IEEE Budapest Power Tech’99 Conference, Budapest, Hungary, Aug 29
– Sept 2, 1999.
[9] IEEE Power System Relaying Committee Report, "Single phase tripping
and auto reclosure of transmission lines", IEEE Trans. Power Delivery,
vol. 7, no. 1, pp. 182-192, Jan. 1992.
[10] E. W. Kimbark, "Selective-pole switching of long double-circuit EHV
lines", IEEE Trans. Power App. & Systems, vol. PAS-95, no. 1, pp. 219-
230, Jan./Feb. 1976.

VIII. BIOGRAPHIES
Mustafa Kizilcay (M’94) was born in Bursa,
Turkey in 1955. He received the B.Sc. degree
from Middle East Technical University of
Ankara in 1979, Dipl.-ing. degree and Ph.D.
degree from University of Hanover, Germany in
1985 and 1991. From 1991 until 1994, he was as
System Analyst with Lahmeyer International in
Frankfurt, Germany. Currently, he is Full
Professor for Power Systems at Fachhochschule
Osnabruck, Germany. Dr. Kizilcay is winner of
literature prize of Power Engineering Society of
German Electroengineers Association (ETG-
VDE) in 1994. He is a Member of IEEE,
CIGRE, VDE and VDI in Germany.
Gábor Bán (M’87. F’94) was born in
Kiskunhalas, Hungary in 1926. He received the
M.Sc.degree from the Budapest University of
Technology and Economics in 1950, PhD degree
(1960) and DSc degree (1980) from the
Hungarian Academy of Science (HAS). His
employment experience included the Budapest
Electric Company, the Electric Power Research
Institute, Budapest. Since 1968 he has been a
professor of the Budapest University of
Technology and Economics. To Dr. Bán were
awarded the Price of the HAS (1970), the
Hungarian State Price (1980). He became the Eötvös Laureate of the HAS in
1997. He is a Fellow of IEEE and the Hungarian Academy of Engineers, a
Distinguished Member of CIGRÉ and a member of the Hungarian
Electrotechnical Association.
László Prikler (M'92) was born in
Mosonmagyaróvár, Hungary on March 7, 1962.
He received the M.Sc degree in Electrical
Engineering from the Technical University of
Budapest in 1986. Following graduation he
joined the academic staff of the Department of
Electric Power Systems of the same University.
His main research interest is computer simulation
of power system transients. In parallel with his
academic carrier he is working for as consultant
and managing director of his own enterprise
Systran Engi. Services Ltd. Mr. Prikler is a
Member of IEEE Power Engineering Society and the Hungarian
Electrotechnical Association. Mr. Prikler was awarded by the Chapter
Regional Outstanding Engineer Award by the Region 8 of IEEE in 2000. He
was the publication committee chairman of Budapest PowerTech’99 and
chairman of IPST’99.
Péter Handl was born in Pécs, Hungary, on July
11, 1977. He studied at the Budapest University
of Technology and Economics. After receiving
the MSc degree in 2000 he is a postgraduate
student at the Department of Electric Power
Engineering. His main fields of interest are
transients in the high voltage power systems,
specially the transients caused by lightning
flashes and the secondary arcs. He is a holder of
a scholarship at the National Power Line
Company in Hungary.
View publication statsView publication stats

Interaction of the secondary arc

More Related Content

What's hot (19)

Similar to Interaction of the secondary arc (20)

Recently uploaded (20)

Interaction of the secondary arc