Capacitive Current Switching Duties of High-Voltage Circuit Breakers:

Background and Practice of New IEC Requirements

R.P.P.Smeets, member, lEEE

A.G.A. Lathouwers
KEMA High-Power Laboratory
Amhem, the Netherlands

Abstract: Capacitive current is switched in power networks charge of the stray- and load capacitances over the
in a number of situations, such as the charging current of system’s inductance(s). This causes high-frequency
overhead lines, cables, GIs installations or the load currents of restrike currents in the system. The flow of the restrike
capacitor- and lilter banks. In each of these situations, the current can - under circumstances - cause unwanted over-
current to be intempted is small or very small. Nevertheless this voltages in the system, due to the inherent high values of
interruption duty has to be considered with the utmost care. di/dt and the associated induced voltages.
Above all, this is due to the inherent risk of re-establishementof In addition, restrike can damage the breaker’s interior
arcing (‘restrike’) because of severe voltage stress after inter-
ruption. Restrike may cause high overvoltages depending on the (e.g. puncture of the nozzle) due to the formation of a new
network topology and can lead to damage to circuit components. arc path not always properly controlled by the cooling
This concern is reflected in a major revision of the relevant IEC activity. After interruption of restrike current, in except-
standard (IEC60056)that is presently under consideration. In the ional cases a renewed situation of trapped charge of high-
revised edition, the concept of a ‘restrike-free’circuit breaker is er value than before, can initiate voltage escalation [l].
abandoned and the regime of testing is sharpened.
load voltap
In the present contribution, the background of the coming .*- *..
IEC60056 standard will be highlighted.Test practice in the high-
power laboratory will be illuminated.
Keywords: Type tests, circuit breaker, capacitive current, capa-
citor bank, SF6. restrike

I. CAPACITIVE LOAD SWITCHING

A. Capacitive load de-energization

Capacitive currents have to be switched in high-voltage
networks in a number of situations. These situations all vciqr over cb\\ i
have in common that, due to the 90 degree lagging load ‘L..,. ./
voltage. at interruption of the power frequency current mulrqt“rpaaronb
the load capacitance is left after at the system peak volt-
age (in the single phase case). Half a period of power
6equency later, the source side ac voltage has become
opposite to this trapped voltage, causing a 2 pu dielectric
voltage stress on the circuit breaker. This is outlined in
-
fig. 1 illustrating the well-known “1 cos” recovery volt-
age.

Breakdown of the circuit breaker contact gap near this

moment of maximum dielectric stress (or any moment
later than one quarter of a power frequency period after
current interruption) is called “restrike” and causes a dis-

Thus, the small interrupted current is rather a disadvant-

age than an advantage. In short-circuit- interruption a
much larger contact distance is needed for interruption,
allowing the transient recovery voltage (TRV) to develop
over fully opened contacts.

0-7803-5935-6/00/$10.00(c) 2000 IEEE 2123

 B. Capacitive load energization
Energization of capacitive load is usually associated with
‘‘inrush‘’ current, the charging current of the capacitive
load over the inductance of the system. Depending on the
nature of the load (capacitor bank, overhead line, cable)
the inrush current has an oscillatory character with a
frequency (in the case of capacitor bank switching) up to
a hundreds of times the power frequency current with an
amplitude of up to few tens of kkpcal. Since this current at
least partially flows during the period of prestrike - i.e.
when an arc is established between approaching contacts -
damage to circuit breakers (andor capacitors) is a risk. Fig. 2 Circuit diagram for voltage jump

For SF6 breakers, the sudden release of energy may da- Upon dropping of the capacitive load C1,the voltage at the
mage the breaker (as with restrike), whereas for vacuum source side of the circuit breaker will “jump” in an
circuit breakers, contacts may weld under the action of the oscillatory manner to a Iower value. It can easily be deri-
high current arc between closing contacts. Such a weld, ved that the relative voltage jump (AUAJ) is given by:
when broken at the next opening, leaves metallic ,with & the capacitive current and I,
-=c the local
microprotrusions, possibly reducing the dielectric strength U 1,
of the gap. especially when the breaking of the weld is not short circuit current as governed by L,. This voltage jump
followed by the annihilation of the protrusions by suffi- manifests itself as an initial TRV and is very important in
cient arcing activity. reducing the probability of restrike, as explained with the
help of fig. 3.
For the energization of overhead lines and cables, the in- In this figure,the dielectrical recovery of the breaker gap
rush current is naturally limited by their surge impedance is assumed to be a straight line for the sake of explanation
(hundreds of ohms for overhead lines and several tens of only. Depending on the moment of contact separation (i.e.
ohms for cables) so that inrush current in these cases is the moment of start of recovery) the following possiblities
usually much smaller than with capacitor banks. arise: the intersection of this line (c in fig. 3) with the
voltage jump oscillation leads to a harmless continuation
Two situations are distinguished here: of power frequency current (reignition), whereas the
0 the “single bank” case where only one single bank is intersection with the 1 - cos recovery voltage initiates a
connected to the bus, where a relatively large source full restrike (b); contact separation at point (a) leads to
inductance limits both amplitude and frequency; complete recovery. Point d is the moment of current inter-
the “back-to-back” case, where a capacitor is con- ruption. From this, it is clear that a higher voltage jump
nected to a bus that already feeds (an)other capacitor (pertinent to a “weak“ location in the grid i.e. low short-
bank@); the small stray inductance gives high fre- circuit current) reduces the probability of restrike.
quency and amplitude. Mutatis mutandis, capacitive switching at strong network
locations is associated with higher probability of restrike.
II. NETWORK CONSIDERATIONS
III. NEW IEC REQlJIFlEMENTS
There are two reasons to consider capacitor bank current
switching as the most severe capacitive switching duty: Since 1987, the existing IEC 60056 is in operation, that -
among others - gives requirements that breakers have to
0 the amplitude of the inrush current is generally much fulfil when assigned a capacitive switching duty [2].
higher (and more detrimental to the breaker) than A major problem in quality assurance of circuit breakers
with distributed components such as lines and cables; is the verification of the absence of restrike events
the frequency of switching of (shunt) capacitor banks throughout their life, based on a limited number of tests.
is much higher than that of lines and cables. This is This especially applies to breakers assigned to capacitor
because shunt capacitor banks are more often than bank switching duties, having an expected number of
not used for the supply of reactive power for reasons switching operations much higher than can ever be tested
of voltage stability at the point of application. Be- economically.
cause the reactive power requirements of the system Therefore, it is rather peculiar that after passing a limited
vary according to the daily load cycle, switching of number of tests (40 threephase and 48 single phase)
shunt capacitors is often on a daily basis, unlike the without restrike a breaker is declared “restrike free”, as
switching of sections of the T&D grid. defined in the present IEC 60056.

A very important parameter is the location of the capaci- In the ongoing revision of the standard, new requirements
tive load. In order to understand this, the simple equiva- for capacitive switching duties are put forward, that are
lent circuit of fig. 2 is of help. believed to be more in conformity with the reality in

(c) 2000 IEEE

0-7803-5935-6/00/$10.00 2124
 is often used, giving values for the voltage jump often
being unrealistically high.
Therefore, one should welcome the revised requirements
giving a more strict and a more workable prescription of
the voltage jump in test laboratories.
In addition, on can have doubts on the necessity of testing
at maximum short circuit current level. Recent infor-
mation on existing stresses on circuit breakers revealed
that the expected maximum short-circuit current in 90%
of the cases is less than 75% of the rated circuit breaker
short-circuitcurrent [4].

a b c d

Wg. 3: Voltagejump, reignition, restrike and m v e r y at capacitive

current interruption

power supply circuits. The concept of a “restrike-free”

breaker is now abandoned, and replaced by two classes:
0 C1: Breakers that are expected to have a low proba-
bility of restrike after passing the tests. This class was
modelled after the existing ANSI C37.09 standard.
0 C2: Circuit breakers that are expected to have a very Table 1: Comparison of numbers of tests existing (grayed) and draft IEC
low probability of restrike after passing the tests 60056 for capacitive switching duties. See notes below.
These new test requirements are more severe compared to
the existing IEC standard in terms of number of tests and Colomn P single (1) m three-phase (3) tests;
Column S: required operating sequence;
preconditioning, the latter for class C2 only. Precondition- Columns jump 2%: tests must be done with a supply circuit the
ing of the breaker (simulatingageing) is prescribed to be a short-circuitimpedance of which is designed to cause S 2% voltage
short-circuit current stress of 60% of the rated short- jump with 10 - 40%of the rated capacitive breaking current
circuit breaking current with rated operation sequence Columns jump 5%: as columns jump 2% but with 5 5% voltage
jump at 100% of the rated capacitive breaking current
(usually 0-CO-CO) prior to the capacitive switching tests. Column 100% 1: tests must be done with a circuit able to supply 5
A comparative summary of the required number of make- 100% of the breakers’ rated shoa-circuit breaking current. Half of
break test is shown in table 1. In the new IEC 60056, the the test arc with reduced ( 2 0 4 % ) of the capacitive current.
rated capacitive breaking current varies from 10 - 500 A Column 10%I: as column 100% I, with 10%Short-circuit current
for line and cable switching (3.6 - 550 kV) and is 400 A In the existing practice no restrike is accepted, in the revised IEC
dmtl, one restrike is allowed if a repetition of the complete series,
for capacitor bank duties for all voltages. When a breaker after finishing of the first series, is free from restrikes.
is assigned to have a rated back-to-back capacitor bank 0 In the new IEC. 2 25%of all tests have to be performed at mini-
inrush current, this current should have a peak value of 20 mum arcing time. No such requirement is in the present IEC 60056.
kA and a frequency of 4.25 kHz . 0 Single phase testing requires a test-voltage equal to the rated phase-
to-ground voltage multiplied by a factor between 1.0 1.7, -
depending on system grounding, type of fault, mutual pbase
As can be seen in table 1. the definition of voltage jump is influence etc.
changed in the revision process. In the existing IEC
60056, the rated short-circuit breaking current ( I,
) is Regarding capacitor bank switching practice, a world
taken as a measure: two supply circuits have to be pre- wide survey was carried out by CIGRE Working Group
pared, one supplying I, and one supplying 0.1 I,. This
13.04comprising43 utilities in 17 countries [3].
In fig. 4, the results on the short-circuit current are sum-
has the disadvantage that the 100% circuit is very difficult
marized. For capacitor banks installed in the weaker part
to realize for rated voltages > 72 kV in most test labo-
of the system, it is reasonable to assume a value in be-
ratories, forcing test engineers to search for a way-out tween the minimum and maximum short circuit power
possibly leading to ill-defined and ad hoc solutions not curve of fig. 4, yielding a 90% percentile around 30 kA.
contributingto the uniformity of testing that IEC pursues.
Practice has established that “the highest available power” This implies, that in 90% of the surveyed cases, the short
circuit current is smaller than 30 U.

(c) 2000 IEEE

0-7803-5935-6/00/$10.00 2125
 1-
common practice.

0.8 .. Another important limiting factor is the available reactive

power. The availability of huge capacitor banks to si-
jO.6.. mulate the capacitive load is a prerequisite not only for
the necessary MVAr of the load, but is also necessary for
4 -.
i0.4
the supply of inrush current when a circuit breaker has the
back-to-back switching duty.
In back-to-back switching, two (large) capacitor banks
must be available at either terminal of the breaker. Since
02 -- CO operation is mandatory, one bank (at load side) of the
I breaker controls the assigned capacitive current, whereas
: : : : : : : : : : the series combination of both banks, together with the
0 10 20 30 40 50 80 (parasitic) inductance of the loop determines both inrush
ylon dmlc c
um(vy
current frequency and peak value. In some cases, the
required small value of this inductance (typically few
pig. 4: Cumulative distribution of short circuitpow- from CERE hundreds of pH) is a problem, such as in situations where
enquiry on capacitive switching practice. From 131.
the banks are physically remote from the testobject. Then,
it may be difficult to realize both frequency and peak
IV.TESTING value of the inrush current. Especially for the lower
voltage ratings (c36 kv) the required surge impedance of
Already long before the new IEC 60056 comes into effect the inrush circuit may be loo low to achieve. In these
(expected in the first half of 2000). testing experience was cases, a solution can be a low-inductive connection to the
gathered at KEMA regarding the revised requirements. In capacitor banks, such as a cable, although this solution
a period of 18 months, 37 tests were requested from cai only be applied up to a limitedtest voltage.
various customers offering switchgear from 72.5 550 -
kV for capacitive tests. 300001
In table 2, the results of these tests are compared for the t
tests according to the existing IEC 60056 (14 objects) and
the tests according to the revised IEC 60056/CDVr e IMX)O 1 available 3-phase power at KEMA

quirements as summarized in table 2.

fail pass
17% 83%
revised IEC 60056/CDV 64% 36%
table 2

The vast majority of customers were testing aimed at

‘very low probability of restrike’. From the results it a p L
pears that the new requirements are definitely more severe 0 CeMeSg +
0
with regard to capacitive current switching.
a +
a. Direct testing A #
A #
In order to realize a sufficiently small (c 5%) voltage IO : .#
jump as defined in the forthcoming IEC standard, supply
circuits in test laboratories must be designed for sufficient ++
short-circuit current at all voltage levels. This implies, &IA
that capacitive current switching tests must be performed
&I ratedvoitageofdrcultbreaker(kV)
in high-power laboratories with sufficient installed power. 1 . ’ ‘ l l e ’
An impression of the requirement of three-phase power 3 10 100 lo00
for direct three-phase testing in conformity with the
revised IEC 60056 for various capacitive switching duties Rg. 5: Required three-phasepower for full pole direct capacitive
switching tests according the forthcomingIEC 60056
can be obtained from fig. 5. Indicated are the limitations
in power and voltage (for full pole three-phase direct test)
of KEMA’s high-power laboratory. Beyond the indicated With 200 MVAr (50 Hz) I 240 MVAr (60 Hz) of
boundaries, tests can be performed on a single phase basis available three-phase reactive power, KEMA high power
or as half pole tests. It must be realized that three-phase laboratory is able to perform full pole direct three-phase
tests are essential for three-phase enclosed switchgear. capacitive switching tests up to 245 kV. Above that volt-
which is at present available up to around 300 kV. For age, single phase tests are possible (KEMA’s single phase
higher voltage ratings, single phase enclosed switchgear is limit 335 MVAd60 Hz and 280 M Y A r b O Hz)up to 420

(c) 2000 lEEE

0-7803-5935-6/00/$10.00 2126
 kV full pole. necessary short-circuit power is not available, synthetic
test methods are a well-known practice. This is a combi-
An example of this is a recent project at KEMA, invol- nation of a (medium voltage) source providing the rated
ving the test of 420 kV circuit breakers (of three different current (not necessarily capacitive) and a high-voltage
types) for back-to-back capacitor bank switching in con- source with restricted current capabilities. The voltage
formity with the revised E C 60056 (class C2 restrike jump can be realized with separate RLC elements [SI. An
probability). All breakers were preconditioned with 37 kA auxiliary circuit breaker separates the current circuit (with
current (3x0) and endured 20 % inrush current at restricted voltage withstand ability) from the HV source.
4.25 kHz at least 120 times. The results are summarized Recently, synthetic tests of a single-break 550 kV 63 kA
in table 3: breaker were succesfully performed at KEMA.
Generally speaking, (capacitive) synthetic testing raises a
number of considerations with respect to the equivalence
of the testing conditions and the network situation:
1. A difficulty with most synthetic test schemes is that
#of CO's testedat 90 A 48 48 the ac and dc components of the recovery voltage are
normally applied at one terminal of the test object, the
damage notin- darnageto other terminal being earthed. In particular for GIs and
obsavation interior spccted interior UHV breakers this is an undesirable situation. Also, in
(PUnCm)
a number of synthetic test circuits (the so called "os-
Table 3: Results of capacitive tests of 420 kV circuit breakers cillatory circuits"), the presence of dc voltage at load
side of the testabject before testing is a disadvantage.
The restrikes (object X, Y) andor the clear visual damage 2. In the case that the recovtry voltage is applied by an
(X.Z)of the interior after inspection convinced the manu- oscillating circuit, a circuit with a very high quality
facturers to modify the design of breakers. These might factor is needed to reach a sufficiently high recovery
very well have passed the more relaxed requirements of voltage. All too often, voltage decay is too rapid.
the existing IEC 60056. An oscillogram with a restrike of 3. A principle disadvantage of synthetic testing is that
this test series is shown in fig. 6. The damage to object Z, the events after a restrike cannot be represented
free of restrikes, is probably caused by the inrush current. correctly. This also applies to the making- and back-
From the above example, it is clear that the combined to-back tests
efforts of manufacturer, utilities, standardizing body and 4. Interaction between circuit and circuit breaker arc can
test authorities will lead to a higher reliability of operation be different in a number of c a w :
of the networks. i) Current chopping (sudden arc extinction of small -
-
e.g. line charging current prior to natural zero) is
b. Synthetic testing known to be influenced by different parasitic elements
In the case that full rated capacitive load andor the and the increased ratio of arc voltage to driving
K.EhlA4 voltage inherent to
synthetic testing. usually,
I the tendency for an arc to
chop is increased in
synthetic test circuits.
ii) Rapid changes in arc
voltage, such as the ex-
tinction peak occurring
prior to current zero in
gas circuit breakers can
introduce strong current
oscillations in capacitive
current [6]. It can easily
be derived that a circuit
like in fig. 2 responds to a
rise of arc voltage with a
rate dddt = S, with a har-
monic current i(t):
i(t) = S,C~[l-cos(ax)J,
with o the oscillation fre
quency of the short-
circuit reactance L)and
6:*i- sampled at 1 0 ldIz Of reStrfking 420 kV breaks undagohg bCk-tO-Wk capacitlvc tesf the load capacitance (CJ.
in conformity with new IEC 60056 (oscts:HF rccordiugs (1 MHz sampling frequeacy) of inrush current
(bottom left) and restrike(top right). Vertical scale of insaa. 8 W d i v .
This current is superim-
posed on the (small) pow-
(c) 2000 IEEE
0-7803-5935-6/00/$10.00 2127
 er frequency current and causes current interruption prior restrike of adequately tested circuit breakers in spite of
to the prospective (i.e. without arc voltage influence) the fact that the (erronous) concept of “restrike-free”cir-
current zero. Because of the different current slope to cuit breaker is abandoned.
zero, the voltage jump is strongly affected by the The higher level of reliability is largely thanks to the
superimposed current, such as is shown in fig. 7, showing increased number of tests and the artificial ageing by
the effect of arc voltage transients on on the voltage jump. preconditioning the breaker. The updated, more robust re-
A case was simulated of 300 A no load line-charging cur- quirements will lead to a larger uniformity, since utilities
rent in a 20 kA circuit at 245 kV. Close to current zero, an will tend to adopt the new IEC requirements, rather than
arc voltage transient is generated: (a) constant arc voltage to stick to own specifications [7]. New explicite and
only; (b) slow arc voltage rise (S, = 8 V/p); (c) steep arc realistic definitions of the initial TRV (voltage jump) will
voltage rise (initial slope 80 V/ps). limit arbitrariness and too easy an escape from test
From this example, it is clear that current distortion due requirements.
As before, preference must be given to direct testing,
10 being the closest match to the situation in real networks.
incremental arc voltage (kv)
When applying synthetic test methods, a number of con-
siderations must be kept in mind, such as correct mag-
nitude of recovery voltage, dielectric stress in GIs and
correct representationof arc-circuit interaction.

VI.REFERENCES
-
[I] CERE WG 13.04, “Capacitive Current Switching State of the
Arf’, Electm No. 155,1994,pp. 33-63
[2] IEC 56, “High-voltage alternating circuit-breakers”, International
Electrotechnical Commission, 1987
[3] CIGRE WG 13.04,“Shunt Capacitor Switching, Sfresses and Test
Methods”, Elecrra. No. 183,1999,13 43 -
[4]A.J.L. Janssen, W. Lanz, “An International Survey on Electrical
Stresses on High-VoltageCircuit-Breakers”, Report on behalf of CIGRE
WG 13.08,CIGRE Conference,Paris, 1998
[5] CIGRE WG 13.04,“Requirements for capacitive c m t switching
20 tests employing synthetic test circuits for circuit-breakers without shunt
16
initialTRV (voltage jump) (kv) -
resisstors, EIecfra87,1983.25 39
[6] G. Daigneault, G. St-Jean, M. Landry, “Comparing Direct and
Synthetic Tests for Interruption of Line-Charging Capacitive Current”,
IEEE Trans. on Power Del. PE-185-PWRDO-03-1997
[7]J.F.Reid, Y.K.Tong, M.A. Waldron, “Contmlled Switching h u e s
and The National Grid Company’s Experiences of Switching Shunt Ca-
pacitor Banks and Shunt Reactors”, CZGRE Conference. 1998,13-112

time (0.1 msldhr) VI1 BIOGRAPHIES

fig. 7:Effect of arc voltage transients (a: none; b gradual; c: steep) R e d Peter Paul Smeets was bom in Venlo. the Netherlands in 1955.
on arc current and initial TRV (voltagejump) He received the M.Sc. degree in physics
from the Eidhoven Univ. of Technology.
to arc voltage transients can have a strong bearing on volt- He obtained a Ph.D. degree for research
age jump and thus on restrike probability. This reasoning work on vacuum atcs. Until 1995,he was
especially applies to cases of small current interruption, an assistant professor at Eindhovea Uni-
versity. During 1991 he worked with Tas-
lower voltages (larger equivalent capacitances) and/or hiba Corporation’s Heavy Apparatus En-
circuit breaker types having high extinction peaks (e.g. air gineering Lahatory in Japan. In 1995,
blast types or radially blown SF6 types). The arc-circuit he joined KJZMA High Power Lab. He is
interaction described above, is different in the case of a “ber of CIGRE WG 13.04, the
“Current Zero Club”. IEEEand CIGRE.
synthetic testing at reduced M V k , with (much) less than
the full rated capacitance, the effect of arc voltage
transients on the initial recovery is much less pronounced.
Andre G.A. Lathouwens received his M.Sc degree in Electrical Eng.
from the bdhoven Univ. of Technology 1988.Rom 1988 to 1990 he
V. CONCLUSIONS worked at “No’s physics and Electronics
Laboratory on calibration procedures for
It is often thought that short-circuit current breaking is the EMP measurements. In 1990 he joined
only severe switching duty. It spite of the much smaller KEMA as a Test Engineer. As a Senior
Test Engineer, he has gained extensive
current to be interrupted, the severity of various capaci- experience in performing type tests on all
tive switching duties must not be underestimated, largely kind of components for electric power
due to the probability of restrike. transmission and distributioa systems.He
The new IEC 60056 requirements with respect to capa- is member of the Short-circuit Testing
Liaison (STL)Technical Committee.
citive current switching will lead to a lower probability of

(c) 2000 IEEE

0-7803-5935-6/00/$10.00 2128
 本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP 图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具