Fourth Workshop & Conference on EHV Technology

CSIC Auditorium, Indian Institute of Science, Bangalore 560 012 India.

WORKSHOP : 15 & 16 July 1998

CURRENT LIMITERS - STATE OF THE ART

Michael Steurer, Klaus Fröhlich
Swiss Federal Institute of Technology Zurich
Zurich, Switzerland

ABSTRACT medium voltage range. However, these systems either

lack on limiting performance or they do not cover the
Discussing the state of the art of current limitation in entire power range needed. Although there are con-
medium voltage range shows the limits and drawbacks tinuous research projects in the field of CLD develop-
of today's current limiting techniques. ment. For the time being none of these approaches led
Focusing on up-to-date research projects in the field, to commercially acceptable systems /2/. Even though
the major part of this paper explains various kinds of there always has been the desire for current limiters /3/,
current limitation by means of super conducting mate- especially the discovery of the so called high tempera-
rials. The two major principles, resistive and inductive ture superconductors (HTSC) with their non linear u-i
limiters, are introduced as well as hybrid approaches. characteristic available at the temperature of liquid ni-
Unsolved problems using high temperature supercon- trogen (T ò 77 K) in 1986 started new efforts to de-
ductors HTSC for current limiting devices are dis- velop CLD's /4/.
cussed. Beside the HTSC projects an alternative ap- This paper starts with the basic considerations on fault
proach, the project "NST - New Switchgear Tech- current limiters, explains why solutions used for low
nology" at the ETH Zurich (Switzerland) is introduced. voltage range can not be scaled to medium voltage and
Concluding from the discussed projects, a commercial describes some of the various forms of novel ap-
realizable fault current limiter for medium (high) voltage proaches for CLD's, mainly those with HTSC.
range based on superconducting materials is not in Not covered in this paper are single phase to ground
sight in the near future. faults in reactor earthed networks with reactors for
short circuit current limitation. This technique is well
known and widely used. Also not addressed are at-
1 INTRODUCTION tempts for novel current limiting systems for un-
The consequences of inevitable fault currents iF in grounded power distribution systems /5/
electric power networks, more than an order of mag-
nitude higher than the nominal current, usually means
severe stress for the affected apparatus such as 2 SHORT CIRCUIT CURRENT LIMITATION

• thermal stress proportional to ∫i F ⋅ dt 2.1 Basics

Figure 1 shows a simple equivalent circuit for dis-
• mechanical stress proportional to ∫ i 2F ⋅ dt cussing the difficulties at short circuit current limitation
in electric power networks. Independent of the load
• damage due to power dissipation at the fault loca- flow prior to the fault, the short circuit current (SCC) iS
tion increases with a certain rate of rise depending on the
circuit parameters (U0 and ZS = RS+jXS) and the phase
Continuously increasing electric power production,
angle of fault occurrence.
distributed with high density meshes, may drive power
networks to the limits of their short circuit current ca-
pability. Novel apparatus such as superconducting
generators, motors, and power lines and the increasing
demand on power quality makes effective short circuit
current limitation desirable /1/.
Many investigations have been carried out so far in the
field of current limitation devices (CLD's), but still only
few systems are commercially available, especially in
 range. Commercial solutions for current limitation in
RS LS iF MV will also be compared.

2.2 Low voltage range

CB
Fault From the equivalent circuit in Figure 1 it easily can be
U0 seen that an inserted voltage drop in the order of mag-
nitude of the source voltage U0 will be sufficient to
force the inductive voltage drop uL to be zero or even
negative. When opening the contacts of a circuit
breaker the voltage drop of a free burning arc between
Figure 1 General equivalent short circuit diagram
(framed part is used in subsequent figures)
the opening contact gap would only be several
10 Volts. The state of the art is to increase the arc's
This leads to the current wave form i1 in Figure 2 when power dissipation by cooling it and splitting the arc
no limiting action takes place (prospective SCC). The into series connected subsections (therefore gaining
simplest way to limit this current would be to choose several cathode drops in series). This increases the
an appropriate high source impedance ZS. This is in- overall voltage drop of LV current limiting switchgear
deed the state of the art technique at medium and high significantly. Suitable devices are produced in large
voltage levels. But as this effects nominal load flow as numbers and are well known as
well it can not be the reasonable technical solution for • Fuses:
the future. Without extra limitation a conventional cir- intensive cooling of the melting wire and finally the
cuit breaker CB breaks the current at t3. arc by the surrounding quartz sand, as well as split-
i ting the arc into sub-sections,
î1 i1 prospective short
circuit current • Circuit breakers:
splitting the arc by metal baffle plates.

i2 , i3 let through current Effective current limitation at prospective currents in

iN the range of up to 100 kA limited to several kA can be
nominal î2
î achieved by those well proven techniques.
load current 3
When the line voltage increases over 1 kV it gets more
t1 t2 t3 t and more difficult to design circuit breakers with cur-
rent limiting capability. Only high voltage heavy duty
fuses with nominal currents of several 100 A can be
Figure 2 Typical current waveforms at fault
conditions built for commercial use.

2.3 Medium voltage range

To limit the first current peak î 1 the limiting device
must react within the time interval t1 and restrict the rise The principles of current limitation stated above are
of current
di
at least to zero (or below) at this point basically the same at medium voltage (MV). Typically
dt RS
in MV networks the ratio is less than in LV. As a
(i2, i3). This can only be done by forcing the voltage LS
drop at the circuit's inductive reactance LS to become consequence the inductance LS is typically higher in
zero uL = LS
di
= 0 , which means the need of inserting i$
dt MV networks assuming the same ratio of $F . There-
iN
an appropriate high voltage drop. Such an action
fore the voltage drop that has to be provided by the
(changing the circuit parameters) can only be provided
limiting device has to be over proportionally higher
by a non linear element and leads to the sketched let
compared to that of LV systems.
through currents, depending upon weather the current
is only limited (i2) or also switched off (i3) at t2. Circuit Some few research projects indeed introduce the
breakers that insert the voltage drop of a burning arc switching arc voltage in special designed circuit break-
are the preferred devices at least at low voltage range. ers for current limitation at MV, but the success is
It shall be noticed that to be efficient, the reacting time rather poor (12 kV/1 kA) /6/. As the number of sub-
of a current limiting device (CLD) must be in the range arcs increase with the nominal voltage, it gets more and
of t1 < 1...1,5 ms for power frequency more complicated to achieve the necessary sub divi-
fN = 60 or 50 Hz. sions from the constructive point of view.
For better understanding of the problems the following If the switch itself can not produce enough voltage
two sections describe the usual way of current limita- drop by the means of an arc one might think of trans-
tion in the low voltage (LV) range (household and in- fer the current to an appropriate limiting impedance ZL
dustry applications) and the reasons why this simple (Figure 3). Therefore the transfer switch TS has to
but effective technique fails at medium voltage (MV) commutate the fault current iF within the time interval t1
(comp. Figure 2) to such an impedance and withstand inserting a resistive or inductive impedance short after
the subsequent transient recovery voltage (TRV). fault occurrence.

TS

nominal and fault conditions

Topological
Increase of impedance at

measures
iF • Splitting into sub grids
• Introducing a higher
voltage range

Passive
ZL • Splitting of bus bars

Figure 3 Inserting a limiting impedance in line at • Transformers with

fault occurrence high stray impedance
• Current limiting air coils
The value ZL at power frequency has to be high
enough to limit the current effective and can easily be

fast increase of impedance at fault

Small impedance at nominal load
calculated to 1...20 Ω minimum for typical MV power • High voltage fuses

Apparatus
measures
ratings. From the 3 basic linear electrical elements
(< 1 kA, < 36 kV)
• Resistor • IS-limiter
(< 4 kA, < 36 kV)
• Inductance
Active

• Capacitor
only a capacitor would be sufficient. Because a not
pre-charged capacitor is a short circuit at the time of novel concepts
commutating, the switch TS could perform that task • Semiconductors
even at MV. The rate of rise of the recovery voltage • HTSC
across the opening switch would be limited by the ca- • Hybrid systems
pacitor to acceptable values. But unfortunately this
would require capacitors of several mF and they would Figure 4 Overview on major current limiting
be commercially unacceptable. measures in medium voltage range including novel
concepts
The other two possibilities left, resistive or inductive
impedances therefore have to have a non linear char- Picking out the most powerful current limiting device
acteristic to provide a sufficient time delay for the commercially available, the IS-limiter shall be described
switch to recover. briefly in the following chapter.
A solution without a switch parallel to the impedance
2.3.1 IS-limiter
needs a non linearity of even higher order. Most of the
novel approaches on CLD's using HTSC follow this The IS-limiter basically consists of a ultra fast acting
principle. switch for nominal loads connected in parallel to a
Current limiting devices using one or both of these two heavy duty fuse. To achieve the very short switching
time, a small explosive charge is used to open the main
principles might be called active (changing the circuits
current path. The current is therefore transferred to the
electrical parameters after fault detection) in contrast to
fuse connected in parallel, which finally limits it within
passive limiting measures where the limitation is per-
0,5 ms.
formed by simply increasing the source impedance ZS
as stated in 2.1. The current flowing through the IS-limiter is measured
by a special trigger unit. To achieve shortest reacting
Figure 4 shows the possibilities for current limitation
times not only the current's instantaneous value is
commercially used by various utilities. Besides topo-
measured but also it's rate of rise. All three phases of a
logic attempts, which are long term solutions that
power system are operated independently /7/.
highly effect grid layout, today two solutions are in use
which are commercially acceptable: In Figure 5, the main parts of an IS-limiter are drawn.
The insulating tube (1) houses the cracking-off bridge
• Increasing the grid impedance by transformer design (3) with the blasting charge (2) in the middle. When
or limiter coils triggered, an electric impulse is coupled by the trans-
• Installation of high voltage fuses or IS-limiters former (5) to the potential of the line. The fuse (4) is
actually dimensioned for much lower nominal currents
Novel approaches for current limitation in MV are all and has therefore a very short reaction time.
settled in the active or "switching" category, meaning
they all use one or both of the two main possibilities:
 • current density jC,
• temperature TC or
• magnetic flux density BC
(Figure 6) lead to the two principles pointed out in
chapter 2.3:
• Resistive current limiters
where the SC is in line with the source and load
• Inductive current limiters
where the limiting impedance is magnetically cou-
pled to the line by means of iron cores.
Figure 5 Cross section of an IS-limiter with
functional parts (picture from /7/)
But there are also concepts not using the SC's quench
but it's negligible resistivity below jC, TC and BC. All
Typical fields of installation for the IS-limiter are: those will be described in the following including se-
lected examples of ongoing projects.
• Bus bar connections
jC
• Coupling of auxiliary supplies with the public grid
• In parallel with a current limiting air coil
IS-limiter are widely used to handle over-stress from BC
Super
short circuit currents. Nevertheless they are not reus-
conducting
able: the explosive charge, the main conductor and the
State
fuse has to be exchanged after tripping.
All new concepts of current limiters have to overcome
at least that lack, and this is indeed not a simple task.
Many novel approaches have been made, and are still
TC
in progress. The most important shall be discussed
now. Figure 6 Typical 3D-Diagram of the critical
parameters of a SC

3 NOVEL APPROACHES 3.1.1 Resistive current limiter

3.1 Superconducting current limiters Using the SC in line with the source leads to the resis-
tive CLD where a principal schematic diagram is given
Since superconducting materials have a highly non lin- in Figure 7. A cryostat holds the SC resistor RSC which
ear behaviour they are principally good candidates to is connected straight to the power line by current
build CLD's. Investigating low temperature supercon- leads, specially designed for minimal heat transfer. The
ductors (LTSC) operating at the temperature of liquid load switch LS in series is necessary to save the RSC
helium (4 K) as well as high temperature supercon- from undue high power loss under fault after tripping
ductors (HTSC) with critical temperatures around the and allows a sufficiently short recovery time (1...1,5 s).
boiling point of nitrogen (77 K) many designs for su- A resistive or inductive shunt ZSh might be added for
perconducting CLD's have been presented. Currently thermal relief as well as for upholding a minimum cur-
there are around 20 projects running world wide in this rent flow.
field /4/.
Whereas CLD's using LTSC are still under develop- ZSh Cryostat
ment, most efforts are made to build HTSC CLD's.
The two most important HTSC materials are
• Bismut-Strontium-Calcium-Copper-Oxide (B2212
and B2223) mostly for filaments and iF
RSC LS
• Yittrium-Barium-Copper-Oxide (YBCO123) mostly
for thin film techniques.
Taking advantage of the quench of an SC, the high
increase of resistivity when exceeding one or more of Figure 7 Schematic diagram of a resistive SC-limiter
the critical parameters such as
When the fault current reaches a value equivalent to jC,
quenching of the SC causes a rise of the resistance RSC
and therefore current limitation. With RSC increasing,
power dissipation heats up the SC and leads to
RSC_WARM , the resistance of the heated SC (approx.
room temperature). Values of the resistivity ρSC_WARM
for common SC materials are in the range of
10-4... 10-3 Ωcm which results in long, thin SC designs
to achieve the necessary resistance in the orders of
several Ω for effective limitation in MV. This is actu-
ally the most important problem to solve when de-
signing HTSC resistive CLD's. The heating is not uni-
form along the entire length because of inhomogeneous
regions within the SC material . This results in so
called "hot spots" which destroy the material locally. RSC Cryostat
So the SC has to be shunted by thin conducting films
(e.g. Ag or Au) to smooth the temperature distribution w1
in length. These shunt films also reduce the heated up
resistivity and lead to even longer stripes. LM
Another attempt to overcome the "hot spot" problem w2
i1 RC Lσ
is to spread thin films (several µm thick) of SC on non LS
conductive substrates. A research project from Ger-
many /8/ works with this technique to develop a resis-
tive SC limiter built up of meander shaped thin film
stripes connected in parallel. Today's switching capa- Figure 8 a: Concept of an inductive SC-limiter
bilities are still rather poor in the range of several kVA. (shielded iron core) /11/,
b: schematic of the equivalent circuit
But since this design allows for a very compact limiter
with minimal weight the project is still carried on.
Looking closer to the actual build-up one can see, that
Even though most of the SC limiter projects today are the secondary coil consists of only one winding: a sta-
on HTSC there are still some in progress with LTSC. ple of rings of SC bulk material (typ.: BSCCO). Only
Both, a British project (63 kV/1,25 kA) /9/ as well as these rings are kept at 77 K by the liquid nitrogen in the
one from Japan (6,6 kV/1 kA) /10/ use low inductive cryostat. Both other main parts, the iron core and the
winded coils. The main problem with LTSC is the un- primary copper winding are at room temperature. This
wanted heat transfer into the cryostat by the current is actually one of the great advantages of this concept,
leads. Therefore the current leads are built of HTSC because there are no current leads to the SC and there-
bulk tubes with a comparatively low thermal conduc- fore minimal thermal losses as stated above. The sec-
tivity. ond advantage of the transformer principle is the pos-
The heat transfer throughout the connectors of a resis- sibility of adjusting the necessary SC coil resistance
tive SC limiter is an inherent problem of that principle after quenching by means of the transformer ratio w.
and therefore the inductive limiter (and variants) is a Typically w1 is chosen to be 1 since it's easier to built
potential alternative to the former. low resistive-high current SC rings than long stripes.
Also the "hot spot" problem is easier to overcome
3.1.2 Inductive current limiter with this design. Again, finally a load switch LS has to
interrupt the current to avoid overheating of the SC.
When speaking of inductive SC limiters, basically the
shielded iron core type is meant. Figure 8 shows the The main disadvantage, beside it can't be used for DC
build-up and the electrical equivalent circuit, which is in applications, are the size and weight in the range of a
principal the one of an short circuited transformer. In transformer equivalent to the nominal power of the
normal operation, the overall impedance of the device CLD. Also the normal conducting primary coil leads to
consists of the DC resistance and the stray inductance unwanted power dissipation trough normal operation.
of both, the primary coil and the SC coil. One can say, Nevertheless this system led to prototypes of highest
the SC coil shields the iron core as the axial magnetic power ratings tested so far. In Switzerland field tests
field in such a "long" SC coil is zero due to shielding of an 1,2 MVA limiter have successfully been per-
currents flowing on the outer surface of the SC coil. In formed at a power station /11/. This CLD consists of 3
the case of a fault, the SC quenches and the value RSC limiter coils of approx. 2 m height with SC rings 38 cm
is transferred to the primary side by the square of the in diameter.
transformer ratio w2, with w = w2/w1. The inductive Besides this two basics approaches there are a variety
SC limiter is thus actually a resistive type, but due to of other concepts introducing SC for CLD's /12/. One
the inductive coupling it's known as the inductive type. of them where the SC stays super conducting during
fault conditions shall be described now.
3.1.3 Transductor limiter 3.2 Project NST

The principle of pre-magnetised iron cores can be em- A complete different approach for a CLD without su-
ployed to build a CLD. The principle is old, but with perconductors is investigated at the High Voltage
SC coils power losses in the bias coils can be reduced Laboratory at the ETH Zurich. The basic idea is to
drastically and therefore there are still ongoing projects combine different modules into a hybrid switching
using the transductor /13/. system. Each one of these modules has a certain task
to handle during switching operation to gain synergetic
In Figure 9 a the principal circuit diagram of a DC bi- effects.
ased CLD is drawn. There are two iron cores with
coils, one for each current direction of the AC load As shown in Figure 10 the main parts of these modules are
current iL. The DC SC coils c1 and c2 keep the iron (A) Mechanical contacts to carry the nominal load
cores at a certain point of saturation and therefore current and to be opened by special designed
minimise the overall inductivity of the device. drive within several 100 µs after fault detection
Figure 9 b clarifies this by showing the magnetic char-
acteristics of the primary coils without (1) and with DC (B) High power semiconductor devices forming a
bias (2', 2'') as well as the resulting curve (3). If iL commutating path
reaches values equivalent to points A' or A'' the in- (C) Non-linear limiting path with a PTC resistor.
ductance rises sharply and the current is limited by the
uSWITCH
inductive voltage drop.
The advantage of the concept is the use of DC instead A
of AC for the superconductor. This avoids the AC iF
losses, which are mostly eddy current losses within the
SC material. Furthermore the SC stays superconduct-
B
ing all the time which means there are no problems with
"hot spots" caused by non-uniform power dissipation
during quenching. On the other hand the size of such a C
device has to be approx. twice the size of an equivalent
transformer and this is indeed a major disadvantage.
Figure 10 Main modules of a hybrid current limiter
iDC DC
cryostat For proper design of the key parts of such a system, a
computer model was set up. Simulated current and
2 iron c‘ c“ voltage transient shapes during short circuit current
cores interruption are shown in Figure 11. Because high rise
of current (di/dt) in the first 1 ms after fault occurrence
is critical, a fault at line voltage maximum was chosen
(symmetric fault current, no DC component).
iL
1.5 uSWITCH
u, i [p.u.]

uLINE
1
i PROSPECTIVE

limiting service limiting 0.5

i TROUGH

B 0

2‘ 2“ -0.5

-1

t [ms]
A‘ A“ i -1.5
0 1 2 3 4 5 6 7 8 9 10

1
Figure 11 Current and voltage waveforms at short
circuit current interruption with a hybrid current
limiter obtained from a computer simulation
Figure 9 a: schematic circuit
b: magnetic characteristic of the DC biased CLD As one can see from the figure above, a short reaction
time of the system is very important. A special actuator
Not only superconductors with their demand on cool- system allows high acceleration of the moving contact
ing equipment besides the still unsolved material prob- in the nominal load path. Energy storage for that drive
lems can be used to design fault current limiters. The is kept low due to a novel design of the mechanical
next section introduces one of the present projects for contact area. First tests on an experimental set-up of
CLD's without SC. such a contact system were already performed and
satisfied the expectations. A second redesigned model
is in process of development to verify the requirements /5/ S. Sugimoto, S. Neo, H. Arita, J. Kida, Y. Matsui, T.
for nominal current ranges up to several kA. Yamagiwa
Thyristor controlled ground fault current limiting
Design of the PTC resistor is still in progress, even system for ungrounded power distribution systems
though several different materials seem to be available IEEE Trans. on PD, Vol. 11, No. 2, April 1996, pp.
to build such a device. A linear resistor (without sig- 940-945
nificant temperature dependence of resistivity) could
/6/ Engelman, Neil; Schreurs, Emile; Drugge
also be sufficient if the required non-linear characteris-
Field test results for a multi-shot 12.47 kV fault current
tic of the system can be transferred to the semicon- limiter
ductor path. IEEE Transactions on Power Delivery v 6 n 3 Jul
In the near future, work will include switching tests of 1991 p 1081-1087
the redesigned model of the mechanical contact system /7/ E. Dreimann, V. Grafe, K. H. Hartung
connected to the high power semiconductor devices Schutzeinrichtung zur Begrenzung von
and a limiting resistor in parallel. These tests will be Kurzschlußströmen
performed at currents and voltages in the range of kA ETZ, Vol. 115, Iss. 9, May 1994, p. 492-494
and kV, respectively. /8/ B. Gromoll, G. Ries, W. Schmidt, H. P. Krämer, P.
Kummeth, H.-W. Neumüller, S. Fischer
Resistive Current Limiters with YBCO Films
4 CONCLUDING REMARKS IEEE Trans. on appl. Superconductivity, Vol. 7, No. 2,
June 1997, pp. 828-831
Because it's technical impossible to employ the princi-
ple of low voltage current limiting techniques for me- /9/ T. Verhaege, [Link].
dium voltage levels many projects have been founded Investigations of HV and EHV superconducting fault
world wide to develop current limiters for MV. Most current limiter
of them rely on super conducting materials, both IEEE Trans. on appl. Superconductivity, Vol. 7, No. 2,
June 1997, pp. 1063-1066
HTSC and LTSC because they have a highly non-
linear electric characteristic when coming out of the /10/ T. Hara, T. Okuma, T. Yamamoto, D. Ito, K. Tasaki,
superconducting state. K. Tsurunaga
Development of a new 6.6 kV / 1500 A class SCFCL
For the time being prototypes in the power range of for electric power systems
1 MVA (HTSC) and 100 MVA (LTSC) have been IEEE-Transactions-on-Power-Delivery. vol.8, no.1;
tested successfully. But nevertheless there are many Jan. 1993; p.182-92
problems to be solved before commercial solutions
/11/ W. Paul, M. Lakner, J. Ryhner, P. Unternährer, Th.
will be in sight. Material problems a s well as costs of
Baumann, M. Chen, L. Wiedenhaorn, A. Guérig
the SC CLD's prevent economical attractive apparatus Test of a 1.2 MVA high-Tc superconducting fault
to be built. Even if the majority of the CLD projects current limiter
are in the field of superconductors, concepts without Supercond. Sci. Technol. 10 (1997) pp. 914-918
SC might also be attractive.
/12/ R. F. Giese
Directory of Superconducting Devic Projects Bearing
5 REFERENCES upon the Electric Power Sector
Argonne National Laboratory, November 1997
/1/ R. F. Giese /13/ J. X. Jin, S. X. Dou, H. K. Liu, C. Grantham
Fault Current Limiters - A Second Look Preperation of High Tc Superconducting Coils for
Argonne National Laboratory March 1995 consideration of their use in a prototype fault current
/2/ V. H. Tahiliani, J. W. Porter limiter
Fault Current Limiters, An Overview of EPRI Research IEEE Trans. on appl. Superconductivity, Vol. 7, No. 2,
IEEE Trans. Power Apparatus and Systems, Vol. P - June 1997, pp. 1051-1054
99, No. 5, Sept / Oct 1980, pp. 1964 - 1969
/3/ P. G. Slade, J. L. Wu, E. J. Stacey, W. F. Stubler, R. E. Address of Authors:
Voshal, J. J. Bonk, J. Porter, L. Hong
The Utilities Requirements for a Distribution Fault Swiss Federal Institute of Technology Zurich
Current Limitier ETH-Zentrum, ETL
IEEE Transactions on Power Delivery, Vol. 7, No. 2, Physikstrasse 3
April 1992, p. 507-515 8092 Zurich
/4/ M. Noe, G. Harms, B. R. Oswald Switzerland, Europe
Supraleitende Strombegrenzer in der Energietechnik
ELEKTRIE, Berlin 51 (1997) 11/12 pp. 414-424