Physica C: Superconductivity and its applications 615 (2023) 1354398

Contents lists available at ScienceDirect

Physica C: Superconductivity and its applications

journal homepage: www.elsevier.com/locate/physc

Fault current limiter based on high temperature superconductors

Pascal Tixador
Univ. Grenoble Alpes, CNRS, Grenoble INP, G2Elab-Institut Néel, Grenoble 38000, France

A R T I C L E I N F O A B S T R A C T

Keywords: High Temperature Superconductors (HTS) made the Superconducting Fault Current Limiter (SFCL) a reality in
Superconductor power electric grids. Resistive SFCL, which is the subject of this article, is based on a unique property of su­
Superconducting Fault Current Superconductor perconductors to quench from a non-dissipative to a high dissipative state above a given current. SFCL is an
(SFCL)
attractive response to the poorly solved problem of fault currents in power electric grids and makes it possible to
Quench power electric grids
Fault currents, Second generation tape
enhance the security and quality of electricity, something that matters more than ever. Liquid nitrogen cryo­
REBCO tape genics – which is possible thanks to HTS- makes the cryogenics user friendly. In-field SFCL operations confirm
Hot spot this. Second generation (2 G) HTS materials or Coated Conductors suit well to SFCL. Due to its high interest, SFCL
Normal zone propagation velocity has been the subject of a great deal of research and development. In terms of Technology Readiness Level (TRL) it
Current flow diverter reaches 8 with several in grid test experiences.
Sapphire substrate Here a short history of SFCL is presented with a highlight about Curl 10, Ampacity and SuperOx devices. The
two major locations in the grid are reported: grid coupling and feeder protection. The quench phenomena is
described in a bit more detail before the basic design (length and cross section) is presented. The design is
primarily thermal, considering two extreme operations: fault with very high prospective current (clear short-
circuit) and fault with prospective current close to the minimum runaway current along the length. This oper­
ation is called the hot spot one. It is due to the intrinsic variation of the critical current along the length and to
the low Normal Zone Propagation Velocity (NZPV). The cost of the superconducting conductor is given, showing
no bottleneck for an economic development. All the more it is still possible to lower the cost with some de­
velopments. This article ends with two real advances presently underway: Current Flow Diverter (CFD) and
Sapphire substrate superconducting 2 G tape. CFD aims to increase the NZPV, to still further secure the operation
and to reduce the conductor cross section. The use of sapphire substrates allows to reduce significantly the length
of the design.

1. Introduction becomes very dissipative. Its growing resistance limits the current,
which cannot overstep a known maximum current. This maximum is
Superconducting Fault Current Limiter (SFCL) is an application not lower than the prospective current, the maximum current without lim­
born with High critical Temperature Superconductors (HTS) but they itation. The reduction of the prospective current depends of the SFCL
have made SFCL a reality in electrical grids [1]. There is a full zoo of location. The current is limited but not interrupted like with a fuse. A
SFCL [2] but we will focus only on the resistive SFCL. It is the most switch gear in series should then cut the current. This stops the dissi­
developed one. It is based on a unique property, generally considered as pation of the SFCL and its temperature rise. A high-speed switch gear is
a drawback: the nearly immediate transition from the Superconducting therefore very favorable. Contrary to the fuse that must be replaced the
(SC) state to a dissipative one as soon as the current oversteps the critical SFCL gets back superconducting after a duration called recovery time
one (Ic). The SFCL works a bit like a fuse in an electrical circuit. It is and may be inserted again in the circuit. If the limitation is nearly im­
inserted in series in the circuit to be protected. The fuse rating is about mediate (µs) the recovery takes time, up to minutes. SFCLs recovering
the critical current of the SFCL. As long as the current remains lower under load are possible [3].
than the critical one the SFCL is “transparent” and very little dissipative Fig. 1 shows typical time evolutions of the current and electric field
for the grid (AC losses and current lead losses mainly). The inductance is in case of a routine fugitive fault: a fault which disappears during its
kept also very low thank to a suitable configuration such as bifilar coil or isolation. The rapid rise of the current after the fault is quickly stopped
pancake. As soon as the current oversteps the critical one, the SFCL by the development of the resistance (increase of the electric field in

E-mail address: [email protected].

https://doi.org/10.1016/j.physc.2023.1354398
Received 25 July 2023; Received in revised form 20 November 2023; Accepted 22 November 2023
Available online 23 November 2023
0921-4534/© 2023 Elsevier B.V. All rights reserved.
P. Tixador Physica C: Superconductivity and its applications 615 (2023) 1354398

Fig. 1). The current is limited to a much lower value (630 A) than the
prospective value (about 1500 A). After the fault isolation (330 ms), the
electric field is zero although a current circulates: the SC element
recovered its SC state.
Any FCL brings a real breakthrough in the grids especially under high
voltages since the fault currents are today not limited but only inter­
rupted by a switch gear after a few cycles in AC. In DC (Direct Current)
grids, especially under high voltages the fault current interruption is still
much more complicated since the current does not pass through zero.
Fault current limitation brings a lot of advantages such as a higher
security/resilience and quality of electrical energy: see Section 4. SFCL is
one of the most advanced HTS devices with the highest TRL (Technology
Readiness Levels) [4]. The TRL reaches 8, meaning in deployment phase,
the SFCL being completed and qualified, but not fully commercialized Fig. 2. Typical architecture of a REBCO tape (the dimensions are not to scale).
(TRL 9). Several companies sell SFCL in the world: SFCL are available for
purchase. Nevertheless, there is much room for improvements and even yield in the development stage. The lengths already available (hundreds
breakthroughs using for example sapphire substrate HTS tapes. of meters) overstep what is required for SFCL (about 100 m).
About a dozen of manufacturers supply REBCO tapes commercially
2. High temperature superconductor throughout the world.
The cost is a key parameter for applications. For superconductors it
Shortly after the discovery of HTS the requirement to texture the HTS in general expressed in Euros or Dollars per kiloampere-meter (€/kA-m
elementary grains appeared as a requirement to get high critical current or $/kA-m): it is the cost of 1 meter conductor carrying 1 kA at 77 K in
densities. self-field conditions. This cost (ClSC ) is then:
BiSrCaCuO (bismuth barium copper oxide (BSCCO)) can be textured
by rather classical mechanical processes. This is why the first HTS
l
CostSC
l
CSC = (1)
conductors were BSCCO under the form of Power In Tube (PIT) tapes or Ic (77 K, sf )
wires called first generation (1 G). However, the matrix around the CostSC
l
is the cost per unit length of the tape (30 to 40 €/m for a 4 mm
BSCCO filaments should be silver. In addition to its high intrinsic cost wide tape in 2023). It is now mainly the process, its complexity, its rate
silver is not favorable for SFCLs which require preferably high re­ and its yield. The costs of the raw materials do not play an important
sistivities. Nevertheless, some BSCCO SFL were built and successfully part today. Eq. (1) shows that the increase of the Ic lower the cost ClSC .
tested [5].
In 2023 an order of magnitude of ClSC is 150 to 200 $/kA-m [9].
For REBaCuO (Rare Earth barium copper oxide (REBCO)) the issue to
Studies shows that 10 $/kA-m is possible in case of very large-scale
get alignment was much more difficult. The solution is to epitaxially
production [11]. For comparison, NbTi cost in about 1 $/kA-m at 4 K.
coat REBCO on a textured flat support (tape) so the name of Coated
It explains why it remains the most used superconductor. Reference [9]
Conductor (CC) for this second generation (2 G) HTS conductors. Ref­
states that 50 $/kA-m is the “tipping point” for broad application in the
erences [6–10] give reviews of superconducting conductors. Fig. 2
energy sector.
shows a typical CC architecture. The substrate (30–150 µm) is the sup­
The superconductors may be also elaborated under the form of bulk
port. Very often Hastelloy® a Nickel based super alloy is used but Sap­
materials. That was the first form of HTS. But this implementation
phire could be a very attractive option for SFCL (see 8.2). A stack of
cannot be economically competitive compared to conductors produced
buffer layers (0.1–0.3 µm) enables the optimal REBCO epitaxial coating
in large quantities. Before the emergence of HTS conductors they have
(1–5 µm). A silver cap layer protects and makes the current injection in
been used for FCL especially BSCCO material. Melt Cast Process (MCP)
the superconductor possible. A stabilizer (10 – 500 µm) more or less
BSCCO shows rather high Jc although polycrystalline on a large scale
complicated, conductive or event isolating (see 7.2) may complete the
since the grains remains locally correlated [12]. MCP BSCCO may be
architecture. The stabilizer plays an important part for SFCL. This
shaped in different geometries. A stabilizer (brass, stainless steel…) in
complex stack and the sophisticated process make it expensive in the
very good contact suits to SFCL can be added for SFCL. REBCO bulks
present development stage and for limited production. REBCO is in
were also used [13] but in a less developed way.
addition sensitive to any disorder. This limits the lengths and the process

Fig. 1. Current (i) and electric field (e) versus time for a fugitive fault and a reclosing, zoom at the first times.

2
P. Tixador Physica C: Superconductivity and its applications 615 (2023) 1354398

3. Short SFCL story

The attractivity of the SFCL and the need of FCL in grids led to a lot of
developments since a long time. An ambitious program was launched by
Alstom and EDF in the 80′ to develop a NbTi high voltage SFCL. NbTi is
the most used Low critical Temperature Superconductor (LTS) [10].
Successful tests were carried out [14,15] but the emergence of the HTS
stopped that program. The cooling at 4 K with liquid helium is not
simple for a grid apparatus and leads to prohibitive refrigeration costs.
Helium experiences in addition regular supply issues. The use of liquid
nitrogen compared to liquid helium fully changes the paradigm for
cryogenics. Liquid nitrogen is on the contrary a very cheap industrial
fluid. Liquid nitrogen enables operation temperatures from about 65 K
since it becomes solid at 63 K. 77 K is the boiling point of liquid nitrogen
at atmospheric temperature. The cryogenic penalty decreases a lot with
the temperature. Cryocoolers with an input power of 7.5 kW at room
temperature extract only 1.5 W at 4 K but 320 W at 80 K, so 213 times
more. Fig. 4. Ampacity NEXANS SFCL (10 kV, 2300 A) in Essen, Germany under
The HTS discovery by A. Müller and G. Bednorz was fantastic for operation since March 2014 (Nexans©).
solid state physics but also for superconducting applications. But the
enthusiasm was too large, suggesting that SC devices would almost flood the utility. It showed nevertheless that a big liquid nitrogen storage tank
the market. May 1987 the cover of Time wrote “The superconductivity close to the HTS device is not appropriate for utilities. The fill-up from
revolution” [16]. The revolution did not take place. A lot of time, trucks is a problem for them. Cooling machines on site suit better.
research and development have been required to have HTS conductors Ampacity demonstrates the maturity of SC cables and SFCL and un­
for applications. Since very few years HTS conductors are available derlies successful in field operation over a very long period (7 years).
fitting nearly all the requirements except the cost still too high for some Ampacity built experience by utilities in SFCL.
applications. The activity about HTS conductors is large in the world and The last project presented concerns a high voltage three phase 220
driven in part by HTS fusion machines [17]. kV 1200 A (460 MVA) SFCL fabricated by SuperOx and commissioned in
The strong interest of SFCLs has led to more than 20 SFCLs operating Moscow (Fig. 5) in 2019 [21]. The highlight is here the very high voltage
in real grids in 2017 [18]. (220 kV) showing another significant advance in the industrial devel­
The SFCL history has followed the development of HTS materials. In opment of SFCL for grids, including HV (High Voltage). The SFCL is
the 90′ HTS were available in the form of bulk materials such as Melt more interesting for High Voltage than for Medium Voltage grids. This
Cast Process (MCP) [12]. It was possible to elaborate Bi2Sr2CaCu2O SFCL uses 8 km of REBCO conductor per phase. This is high but we will
(BSCCO 2212) tubes a few centimeters in diameter and up to 0.5 m in see that solutions exist to lower this length. In 2020 this SFCL experi­
height. These tubes were machined to elaborate superconducting ele­ enced four short-circuits and operated flawlessly exactly as designed. It
ments for SFCL. Curl 10 [19] was one of the first significant SFCL project provides uninterrupted power supply to customers since 2019 with
using BSCCO tubes. One bifilar tube 58 mm in diameter developed a readiness of the cryogenic system.
total length of 5.4 m and showed a Ic of 850 A at 65 K (operating tem­ In summary they are successful in field operations of SFCL up to High
perature of the SFCL). 32 tubes were put in series per phase (Fig. 3). Curl Voltages. Their behaviour fully matches their design whereas cryogenics
10 was funded by German BMBF with several partners including Nexans. does not pose particular issues. Cryogenics often is the main concern for
This 3 phase 10 kV – 600 A (10 MVA) SFCL was successfully installed in utilities. SFCL are ready for use in power systems.
March 2004 during one year in the RWE grid at Netphen, Germany. A very few SFCL experiences were nevertheless unsuccessful. The
Nevertheless, it never experienced a severe short-circuit during this field 115 kV SFCL installed in Thailand was damaged during over voltage
test experience. Curl 10 was the first successful experience of a SFCL in a conditions. This only shows that the isolation was not correctly
real grid. designed. Anyway, zero risk does not exist. Transformers still are
As soon as Coated Conductors emerged they have been used in SFCL destroyed today whereas they benefit from over 100 years of industrial
since they suit well. So Nexans switched from BSCCO tubes to 2 G tapes.
A major project is Ampacity [20] in downtown Essen, Germany.
Ampacity includes a 1 km long SC cable together with a SFCL (Fig. 4).
The three phase 10 kV 2300 A (40 MVA) SFCL uses REBCO tapes
operating at 71 K or 77 K when the operating current is 1400 A. The
highlight of this project is its successful operation during 7 years
(2014–2021) without interruption. The SCFL experienced several faults
and operated exactly as designed. Cooling remained nearly invisible for

Fig. 5. SuperOx 3 phase SFCL (220 kV, 1200 A) in Moscow, Russia under
Fig. 3. Curl 10 SFCL - Bi-2212 tube and assembly for the SFCL. operation since 2019 (SuperOx©).

3
P. Tixador Physica C: Superconductivity and its applications 615 (2023) 1354398

experience.
SS/C
A standard FCL test procedure [22] already exists for AC systems peak
Ipros = k √̅̅̅ (2)
above 1 kV. Simulation models of SFCL [23] are accurate and may be 3 Un
inserted in power system software.
Where Un is the RMS phase to phase rated voltage and k the asymmet­
Before to study more in detail the SFCL, let’s see why a FCL is so √̅̅̅
rical coefficient (about 2.55) including the ratio 2 between maximum
interesting in electric grids, what it brings. It will define in addition some
and RMS current values.
of the requirements for the SFCL presented just after.
We will see that such a severe fault is not an issue for the SFCL.
But most of the faults are phase to earth faults or faults through an
4. Contributions of the FCL in electric grids
impedance (tree or animal contacts, lightning…). They show rather low
prospective currents.
The possible locations are numerous et some have been even not yet
Overhead lines experience many faults called transient: they disap­
imagined. We will present here only the two main locations (Fig. 6): bus
pear when the current is cut. As example, birds between two phases
bar / grid coupling and feeder protection. Reference [24] present others.
quickly burn and ashes fall as soon as the line is cleared so that the line
Concerns in electrical grid are to enhance the quality of service and
can be returned to service. This re-close function is most of the time
the supply security. A very simple solution is to multiply the supply ways
automatic. In France the delay between opening and first reclosing is
to the load. But this practice, called meshing, is limited by the pro­
300 ms for overhead lines. On the other hand, for cables the faults are
spective current, which increases: it is the sum of the prospective cur­
persistent and the reclosing is not automatic.
rents of each supply. Meshing is then quickly restrained by the
In summary the fault currents may take any value up to the short-
destructive effects of high currents except if a SFCL is used. A SFCL
circuit current (Eq. (2)). A quick reclosing is often required for conti­
makes indeed the prospective current a theoretical quantity since the
nuity of service.
SFCL guarantees a fault current under a given value nearly indepen­
dently of the prospective one. By managing the fault current, a SFCL
enables a perfect grid with an infinite prospective current. The strong 6. Quench of a SC element
possible meshing thanks to SFCL make the grid very robust and resilient,
two present demands. This is why SFCL are often used for grid coupling. Resistive SFCL are based on the loss of the superconducting state
This connection between the two grids secures the supply of all the when the current oversteps the critical value. But the local critical value
feeders of both grids but is impossible without SFCL since the prospec­ intrinsically varies along the length of the 2 G REBCO tapes. Fig. 8 shows
tive currents are then too high. A least this location reassures the utilities for a THEVA [25] tape the Ic measured every 1.1 mm using the standard
since if the SFCL does not operate, we go back to the initial situation criterion of 100 µV/m at 77 K for a length of 100 m. The minimum, mean
with the two uncoupled grids. The recovery time is not an issue, another and maximum values are respectively 642.7 A, 743.5 A and 785.6 A.
advantage of that location. The SFCL is inserted in an electrical circuit which plays a very
The second location is the protection of feeders. It is more classical. important part. Fig. 9 gives a simple model with a voltage source (rated
Its limits the fault current of the protected feeder and its contribution to RMS value Vn), the grid and fault resistances and inductances (Rgrid, Lgrid,
the grid above. Another advantage is to quickly isolate the fault by Rfault, Lfault) and the SFCL. For a resistive SFCL considered in this article
developing a voltage across the SFCL as illustrate by Fig. 7. In case of a the SFCL is mainly a variable resistance (RSFCL). As seen previously the
bolted short circuit, without SFCL, the other loads are no more supplied: fault impedance varies a lot, from zero (bolted fault) to high values. The
they are short-circuited up to the opening of the switch gear. But with a RMS prospective current in permanent regime (IPR pros ) is:
SFCL after a very short transient the other loads are still supplied by Vn
nearly the rated voltage (Fig. 7). For loads very sensitive to voltage sags IPR
pros = √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
( )2 ( )2̅ (3)
SFCL is highly interesting. Rgrid + Rfault + Lgrid ω + Lfault ω

The fault impedance (Rfault, Lfault) defines two configurations:

5. Specifications for a FCL
voltage and current sources. When the fault impedance is zero or low the
voltage drop across the grid elements (Rgrid, Lgrid) is little and the SFCL is
The faults in an electrical grid are very various. The prospective
supplied by a voltage (Vn). The prospective current is very high but the
currents vary a lot.
SFCL resistance limits the current. On the other hand when the fault
The extreme case is the bolted fault with a zero impedance between
impedance (Rfault, Lfault) is high enough to be larger than the SFCL
all the conductors. The fault currents are maximum. The peak value is
resistance, the current is given. The circuit supplying the SFCL is a
given by the short circuit power (SS/C). The expression for a three phase
current source with the prospective current as RMS value.
AC grid is:
In case of a voltage source (high prospective current) the current
rises very rapidly and quickly oversteps the critical current along the SC
element. A large part of the SC element quenches and dissipates.
Nevertheless, even in this case the SC element does not integrally
quench immediately. Quench starts indeed at several locations and
quickly propagates all the more the current is high. The length, which
initially quenches is proportional to the applied voltage [26]. Decroux
et al. [26] introduced a High Dissipative State (HDS) when the current
reaches about 3 Ic. Their measurements show that the HDS is intrinsic
and occurs under the critical temperature. Propagation in the HDS is
much quicker than the thermal normal zone propagation velocity
(VNZPV) which is low for HTS. A simplified expression of VNZPV in
adiabatic conditions is [27]:
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
J ρλ
V NZPV ≈ (4)
cp T c − T o
Fig. 6. Two possible locations for SFCL: grid coupling (SFCL1) and feeder
protection (SFCL2).

4
P. Tixador Physica C: Superconductivity and its applications 615 (2023) 1354398

Fig. 7. transformer current and load voltage (Fig. 6) with and without SFCL protecting a feeder.

the order of the minimum Ic along the SC element. More precisely it is

the minimum runaway current along the SC element but it is very close
to minimum Ic. At Ic the dissipation is indeed low enough to be absorbed
by the coolant. Considering these currents about Ic there is no HDS and
the NZPV defines the propagation, which is so low (1 m/s) that it can be
neglected.

7. SC element design

The SC element design derives from the behaviors described in the

previous section, from the two operations: high prospective current fault
and prospective current close to Ic fault.
The design of the SC element is basically a thermal design. When the
SC element loses its superconducting state it dissipates during the
clearing time (Δt) up to the opening of the circuit by the switchgear.
The design must keep the maximum temperature (Tmax) reached at
Δt under a value which does not at least damage the tape. The recovery
time, linked to Tmax, may set additional constraints. A maximum tem­
Fig. 8. Ic (77 k, sf) measurement every 1.1 mm along 100 m of THEVA perature between 350 and 400 K is often considered in the designs.
tape [17]. We have seen that they are two extreme regimes:

• limitation regime for bolted fault; the source is a voltage source and
we may assume that the full length very rapidly quenches
• hotspot regime for a prospective current close to the minimum crit­
ical current along the length which is approximately Ic (RMS value);
the source is a current source and only a limited length quenches
possibly even a single hot spot, the resistance of this/these spot(s)
remains little compared to the fault impedance since the NZPV is
low; the current remains nearly constant (RMS value).

To get analytical expressions to extract the relevant parameters some

assumptions are considered:
Fig. 9. electrical circuit with the voltage source the grid and fault resistances
and inductances as well as the SFCL variable resistance.
• adiabatic conditions; the dissipation is much higher compared to
what can extract the coolant (liquid nitrogen)
• isothermal tape/conductor; its cross-section is reduced so that the
J: current density; cp: specific heat per unit volume; ρ: resistivity; λ:
thermal diffusion is good within the clearing time.
thermal conductivity, To, Tc: operating and critical temperatures.
The order of magnitude of the NZPV at the critical current density at
With these two assumptions the thermal equation becomes:
77 K is between 0.1 to 10 m/s [28,29], which is low. Within the clearing
P(T, t) dt = volume cp(T) dT (5)P(T, t) : dissipated power in the volume;
time (typically 50 ms), this represents only 5 mm (1 m/s). We will see
cp : specific heat per unit volume of the conductor.
that the Current Flow Diverter (CFD) concept [30] makes possible to
We will see that a stabilizer has to be bonded to the REBCO tape to
significantly enhance the NZPV (see 8.1).
survive the hot-spot regime. The conductor refers to the tape bonded to
For voltage sources, the hypothesis that the full length becomes
its stabilizer.
resistive is rapidly valid due to HDS.
In current source configuration (low prospective current) the quench
behaviour is different above all if the current amplitude corresponds to

5
P. Tixador Physica C: Superconductivity and its applications 615 (2023) 1354398

7.1. Conductor cost design.

A Hastelloy® substrate shows a high resistivity (1.25 µΩm). To cope
The SC conductor cost is vital for the SFCL. Since the cost of a coated with the hot spot regime a stabilizer is bonded to the REBCO tape (see
tape is expressed in €/kA-m (ClSC ) the cost of the SC element (CSC El )
next section). Considering also a Hastelloy® stabilizer and neglecting
becomes: the other conductive layers the maximum electric field is about 160 V/m
(Δt = 50 ms). This electric field is 3 times higher compared to the value
CSC l
El = CSC Ic (77 K, sf ) LEl (6) for the SuperOx 220 kV SFCL brought back to the same clearing time.
LEl is the SC element length. The operating current (Io) is a fraction of There is a lot of room for improvement and cost reduction. In the context
the critical current at the operating temperature Ic(To, sf). For resistive of FASTGRID project [31] we succeed to reach 130 V/m repetitively and
SFCL the tape operates in self field conditions. This fraction (cgrid) is safely at 65 K with a tape with a thick Hastelloy® stabilizer.
mainly given by grid constraints, an order of magnitude is 0.4. The Another way has been investigated to reach high electric fields under
length (LEl) is given by the operating voltage (Vo) divided by the electric limitation: dielectric (electrical insulator) stabilizers [32,33]. It makes
field (Elim) supported by the conductor under limitation. The cost possible high theoretical electric fields (440 V/m [33]) but this solution
expression becomes: remains very challenging to implement.

ClSC Ic (77 K, sf ) V o Io 7.3. Hot spot regime

CSC
El = (7)
ELim Ic (T o , sf ) cgrid
The power supply is then a current source (RMS value Ic) and only a
Vo, Io and cgrid are given by the grid. The merit factor for the SFCL
very few spots quench and propagate very little. Equation (5) becomes:
cost is then:
( )
Lspot
ClSC Ic (77 K, sf ) RHS i(t)2 dt = Lspot Stot cp dT RHS = ρ (13)
(8) Stot
ELim Ic (T o , sf )
Its minimization to lower the SFCL cost is achieved by: ρ
Lspot
i(t)2 dt = Lspot Stot cp dT (14)
Stot
• a low ClSC (production volume, process yield, Ic performance) Lspot: total length of the hot spot(s).
• a low operating temperature to benefit from higher Ic, at 65 K, the We may introduce the engineering current density (je(t) = i(t)/Stot).
minimum temperature with liquid nitrogen, Ic is doubled compared The integration of this equation gives the maximum temperature in
to 77 K HS
hot spot regime (Tmax ):
• a high electric field in limitation regime Elim.
HS
Tmax
∫Δt ∫
cp
We will get a simplified expression for Elim whereas the hot spot
2
j(t) dt = dT (15)
ρ
regime gives constraints. 0 T0

7.2. Limitation regime For a sinusoidal current density (Jce: RMS value) the hot spot maximum
temperature is given by:
The power supply is in this regime a voltage source and nearly the HS
Tmax
∫
total volume quenches. Equation (5) becomes: cp
2
Jce Δt = dT (16)
( ) ρ
v(t)2 Ltot T0
dt = Ltot Stot cp dT RLim = ρ (9)
RLim Stot
For HTS the critical current per unit width (Ic-w) is more used and
RLim : resistance of the SC element; Ltot, Stot : total length and cross relevant. Introducing Ic-w the equation becomes:
section of the SC element; ρ: equivalent conductor resistivity
We may introduce the electric field: e(t) = v(t)/Ltot. Ic− w
e = √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
∫ HS ̅ (17)
1 T max cp
2 Δt T0 ρ dT
v(t)
2
dt = e(t)2 dt = ρ cp dT (10)
Ltot
e: thickness of the conductor.
Considering the resistivity of Hastelloy® (1.25 µΩm) to get a high
The time integration of this equation up to the clearing time Δt gives the
Lim electric field and Ic-w = 1000 A/cm-w (value at 65 K) the thickness
maximum temperature in limitation regime (Tmax ):
reaches about 0.8 mm. The typical thickness of a tape is under 100 µm. A
∫Δt
Lim
Tmax
∫ stabilizer has to be bonded to the tape to have a safe hot-spot regime.
2
e(t) dt = ρ cp dT (11) Hastelloy® suits since the tape is mainly Hastelloy® (no differential
0 To
contraction) and since it shows a high resistivity. The bonding of such a
thick stabilizer remains chalenging and the conductor becomes stiff.
For a sinusoidal voltage, the integration of the electric field gives the
RMS value of the electric field. 8. Superconducting tapes suitable for SFCL
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
∫ T Lim
max
ρ cp dT Considering the merit factor for the SFCL cost (Eq. (8)) and the
(12)
T0
expression of the electric field under limitation (Eq. (12)), the main
lim
ERMS =
Δt
requirements for a SC tape for SFCL purpose are:
This expression shows that a high resistivity and a short clearing time
enhance Elim
RMS reducing so the SC element cost. • High Ic-w (direct influence on ClSC ; low conductor and SFCL cost)
The clearing time is given by the performance of the switchgear. 20 • High resistivity in normal state (high electric field)
ms is possible with a suitable high-speed switchgear. But we have to • High NZPV (hot spot regime less severe)
consider the case the switchgear fails to open. A second device opens
then the circuit so that a clearing time of 50 ms is often considered in the

6
P. Tixador Physica C: Superconductivity and its applications 615 (2023) 1354398

8.1. CFD progress thickness was moderate. The results were excellent but the very high
cost of the SC elements prevented from any further industrial
The NZPV is poor for conventional REBCO tapes which leads to the development.
hot spot regime. The NZPV enhancement makes the hot spot regime With a sapphire substrate tape the process conditions are much less
much less severe whereas it makes possible to reduce the conductor favorable. For cost reasons the sapphire surface cannot be EPI polished,
(tape + stabilizer) thickness [34]. By modifying the tape architecture, only intermediately polished. The REBCO thickness should be large to
the NZPV can be enhanced by a least one order of magnitude: it is the carry high currents and cracks should be avoided. It remains very
Current Flow Diverter (CFD) concept [30]. challenging when the thickness oversteps a few hundreds of nm. These
The initial idea was to partially isolate electrically the SC layer from two challenges (surface quality and thickness) have been studied and
the metallization one (Ag) to divert the current flow above Ic (Fig. 10). carefully analyzed [44]. The growth conditions have been optimized.
This current diversion increases the local current density and dissipation Thermal annealing has improved the surface quality of intermediately
resulting in a significant enhancement of the NZPV. P. Barusco and Al. polished sapphire. Optimized oxygenation made possible to growth
[35] investigated for example the introduction of a Y2O3 insulating layer REBCO layers without cracks up to 600 to 700 nm.
between the REBCO and Ag layers on about 80% of the surface. Ref. [36] Tel Aviv University succeeded to elaborate a 1 m long sample on
investigates another way using Ag-In intermetallic coating. The intermediate polished single crystal sapphire (continuous pulling
measured NZPV is increased by nearly one order of magnitude. method) [31]. Its critical current was 16 A (77 K, 100 µV/m) on the full
Another concept is b-CFD (buffer layer-CFD, [29,37]). The length. Higher Ic were measured on shorter sample lengths (25 A (77 K,
REBCO/Ag interface is not modified. The change is the thickness of the sf) on 0.55 m for example).
Ag layer: thick on the substrate side and very thin on the REBCO side. Within the FASTGRID project, a sample from Tel Aviv University,
This diverts the current on the edges of the tape increasing the local showed at 65 K an electric field of 1060 VRMS/m for a clearing time of 50
dissipation and boosting the NZPV. P. Barusco and al. [38] partially ms [34]. The calculated mean temperature was only about 260 K
sulfurize the Ag coating on the REBCO (about 1 µm) reducing the pure showing that higher electric field are possible.
Ag thickness to about 0.1 µm. They increase the voltage rise by 50 (60 The today cost of sapphire long tape is very high (500 $ for 1 m, 10
V/s versus 1.2 V/s). mm wide and 1 mm thick) but strong cost reductions are foreseen. The
CFD definitely improves the robustness against hot spots. CFD has increase of the Ic is also mandatory to make this appealing solution
been experimentally validated but the low cost CFD industrial imple­ economically viable.
mentation remains challenging. However advances are underways.
9. Conclusion

8.2. Sapphire substrate tape [40,41] SFCL is an innovative SC application which brings a new attractive
function to electric grids by making them more resilient inter alia. It is
A high resistivity is very favorable for a taped dedicated to SFCL only possible economically with HTS to benefit from simple liquid ni­
since it increases the electric fields under limitation (Eq. (12)). Ref. [41] trogen cryogenics. In addition, REBCO tapes, the most advanced HTS
reports an electric field of 1400 VRMS/m. conductors, suit well. There is no real non SC equivalence at least under
A REBCO tape with a sapphire substrate is then a very appealing high voltages where the demand is. With a TRL 8, SFCL clearly came out
solution. The architecture of the sapphire substrate tapes is in addition of laboratories, as proof the SFCL in service in power grids. One SFCL
rather simple with only two buffer layers (YSZ and CZO) between the was in successful service during 7 years and another flawlessly operates
sapphire and the REBCO layer. The two sides of the substrate can be used in the 220 kV grid of Moscow.
to coat the SC layers. Its NZPV is higher when compared to conventional SFCLs are an attractive response to the growing fault current issue in
tapes since the thermal conductivity of Sapphire is about ten times power grids and up-coming changes associated to the strong increase of
higher when compared to Hastelloy®. Simulations [42] show that CFD the electric vector in the future for carbon-free economies by 2050. In
applied to sapphire substrate tapes, still enhancing the NZPV. France the electrical energy will increase by 30% by 2050 whereas the
Sapphire substrate tapes bring a real breakthrough for SFCL but its total energy will decrease by 40%. The exchanges on the electric grids
development stage remains much less advanced than for Hastelloy® will grow a lot and the SFCL accommodate this grow without having to
substrate tapes. This breakthrough was possible thank to the availability upgrade existing installations, a very costly operation.
of long sapphire tapes. The coating of REBCO on sapphire is indeed not The cost of the SC conductor, especially in the future, should not
new and was used for SFCL by Siemens for example [43]. But it con­ hinder the economic development. The present main brake certainly is
cerned sapphire wafers coated by REBCO and etched to get long me­ the reluctance of the utilities to use a revolutionary device in their grids
anders. Decroux et al. [26] optimized the geometry and the and the fear of cryogenics even if in field experience over long times
metallization layer to improve the operation especially in the hot spot proves the nearly invisibility of this servitude for them.
regime. It dealt with wafers with an EPI polished surface. The REBCO

Fig. 10. diagram of the CFD and b-CFD principles [39].

7
P. Tixador Physica C: Superconductivity and its applications 615 (2023) 1354398

Even if SFCL has reached an advanced development stage it can still Superconductors for fusion: a roadmap, Superconduct. Sci. Technol. 34 (2021),
102001.
experience a lot of improvement and even breakthroughs thanks to the
[18] M. Noe, Superconducting fault current limiters, in: presented at the2017 EUCAS
emergence of sapphire substrate REBCO tape for example. Short Course Power Applications, Geneva, Switzerland, 2017. September 17th.
The prediction of the future market for SFCL is difficult and some­ [19] R. Kreutz, J. Bock, F. Breuer, K.-P. Juengst, M. Kleimaier, H.-U. Klein, D. Krischel,
what risky. But SFCL has clearly matured for use in electric grids M. Noe, R. Steingass, K.-H. Weck, System technology and test of CURL 10, a 10kV,
10 MVA resistive high-Tc superconducting fault current limiter, Trans. Appl.
whereas they bring a satisfying answer to the future needs and evolu­ Superconduct. 15 (2005) 1961–1964.
tions. However, it will be a rather niche market except if the costs (su­ [20] M. Stemmle, F. Merschel, M. Noe, Ampacity project – update on world’s first
perconductor and cryogenics) dramatically decrease. Room temperature superconducting cable and fault current limiter installation in a German City
Centre, in: 23nd International Conference on Electricity Distribution, Lyon, 2015
suitable superconductors will fully change of course also the data. paper 676.
[21] M. Moyzykh, D. Gorbunova, P. Ustyuzhanin, D. Sotnikov, K. Baburin, A. Maklakov,
E. Magommedov, A. Shumkov, A. Telnova, V. Shcherbakov, D. Kumarov,
CRediT authorship contribution statement L. Sabirov, M. Medovik, A. Kadyrbaev, S. Alexandrov, I. Mikoyan, S. Samoilenkov,
A. Vasilov, First Russian 220kV superconducting fault current limiter (SFCL) for
Pascal Tixador: Conceptualization, Writing – original draft, Writing application in city grid, Trans. Appl. Superconduct. 31 (2021), 5601707.
[22] IEEE guide for Fault Current Limiter (FCL) testing of FCLs rated above 1000V AC,
– review & editing. IEEE Std C37.302TM-2015, 2016.
[23] F. Sirois, Modeling Resistive Superconducting Fault Current Limiter in Power
System Transient Simulators, World Scientific Publishing, 2018. ISBN-10:
9813272972, chapter 7.
Declaration of Competing Interest
[24] P. Tixador, Major SFCL Locations with Economical Aspects (2018) 201–235.
DecISBN-10: 981327297X, chapter 10.
The authors declare that they have no known competing financial [25] https://www.theva.com.
[26] M. Decroux, L. Antognazza, S. Reymond, W. Paul, M. Chen, Ø. Fischer, Studies of
interests or personal relationships that could have appeared to influence
YBCO strip lines under voltage pulses: optimization of the design of fault current
the work reported in this paper. limiters, Trans. Appl. Superconduct. 13 (2003) 1988–1991.
[27] M. Wilson, Superconducting Magnets, Clarendon Press, 1987.
Data availability [28] M. Bonura, C. Senatore, Temperature and field dependence of the quench
propagation velocity in industrial REBCO coated conductors, Trans. Appl.
Superconduct. 27 (2017), 6600705.
Data will be made available on request. [29] C. Lacroix, J. Giguère, S.-M. Bergeron Hartman, H.Ben Saad, A. Martin, T. Leduc,
M. Gendron-Paul, Z. Bellil, J.-H. Fournier-Lupien, L. Moret, P. Barusco,
X. Granados, X. Obradors, M. Pekarcikova, F. Gomory, V. Grosse, M. Bauer,
F. Sirois, Normal Zone Propagation in Various REBCO Tape Architectures, 35,
References Superconductor Science and Technology, 2022, 055009.
[30] C. Lacroix, F. Sirois, Concept of a current flow divertor for accelerating the normal
zone propagation velocity in 2G HTS coated conductors, Supercond. Sci. Technol.
[1] P. Tixador, Superconducting Fault Current Limiter: Innovation for the Electric Grid,
27 (2014), 035003.
World Scientific Publishing, 2018. ISBN-10: 9813272972.
[31] P. Tixador, A. Akbar, M. Bauer, M. Bocchi, A. Calleja, C. Creusot, G. Deutscher,
[2] M. Noe, M. Steuer, High-temperature superconductor fault current limiters:
F. Gomory, M. Noe, X. Obradors, M. Pekarcikova, F. Sirois, Some results of the EU
concepts, applications, and development status, Supercond. Sci. Technol. 20
project FASTGRID, IEEE Trans. Appl. Supercond. 32 (5601006) (2022), https://
(2007) R15–R29.
doi.org/10.1109/TASC.2022.3151318. Art no.
[3] A. Berger, M. Noe, A. Kudymow, Test results of 60 kVA current limiting
[32] M. Pekarčíková, J. Mišík, M. Drienovský, J. Krajčovič, M. Vojenčiak, M. Búran,
transformer with full recovery under load, IEEE Trans. Appl. Supercond. 21 (2011)
M. Moša, T. Húlan, M. Skarba, E. Cuninková, F. Gömöry, Composite heat sink
1384–1887.
material for superconducting tape in fault current limiter applications, Materials
[4] B.G. Marchionini, Y. Yamada, L. Martini, H. Ohsaki, High-temperature
(Basel) 13 (2020) 1832.
superconductivity: a roadmap for electric power sector applications, 2015-2030,
[33] P. Tixador, J. Vialle, A. Zampa, A. Badel, Stabilizers for REBCO conductors for high
IEEE Trans. Appl. Supercond. 27 (2017), 0500907.
performance SFCL, IEEE Trans. Appl. Supercond. 33 (2023), 5601110.
[5] L. Martini, M. Bocchi, M. Ascade, A. Valzasina, V. Rossi, C. Ravetta, G. Angeli, Live-
[34] A. Zampa, Experimental and Numerical Study of 2G HTS Tapes to Develop High-
grid installation and field testing of the first Italian superconducting fault current
Performance R-SFCL conductors, PhD (2021).
limiter, IEEE Trans. Appl. Supercond. 23 (2013), 5602504.
[35] P. Barusco, X. Granados, L. Saltarelli, J.-H. Fournier-Lupien, C. Lacroix, H.Ben
[6] D. Larbalestier, A. Gurevich, D. Matthew-Feldmann, A. Polyanskii, High-Tc
Saad, F. Sirois, V.R. Vlad, A. Calleja, V. Grosse, T. Puig, X. Obradors, Chemical
superconducting materials for electric power applications, Nature 414 (2001)
solution deposition of insulating yttria nanolayers as current flow diverter in
368–377.
superconducting GdBa2Cu3O7-δ coated conductors, ACS Omega 7 (2022)
[7] X. Obradors, T. Puig, Coated conductors for power applications: materials
15315–15325.
challenges, Supercond. Sci. Technol. 27 (2014), 044003.
[36] P. Barusco, H. Ben-Saad, D. Horn-Bourque, C. Lacroix, F. Sirois, T. Puig, J.
[8] J.L. MacManus-Driscoll, S.C. Wimbush, Processing and application of high-
Gutiérrez, X. Granados and X. Obradors, Enhanced normal zone propagation
temperature superconducting coated conductors, Nat. Rev. Mater. 6 (2021)
velocity in REBCO coated conductors using an intermetallic stabilizer coating,
587–604.
submitted to IEEE Trans. Appl. Supercond.
[9] A. Molodyk, D. Larbalestier, The prospects of high-temperature superconductors,
[37] C. Lacroix, F. Sirois, J.H. Fournier Lupien, Engineering of second generation HTS
Science 380 (2023) 1220–1222.
coated conductor architecture to enhance the normal zone propagation velocity in
[10] A. Godeke, High temperature superconductors for commercial magnets,
various operating conditions, Supercond. Sci. Technol. 30 (2017), 064004.
Supercond. Sci. Technol. 36 (2023), 113001.
[38] P. Barusco, J. Giguère, C. Lacroix, F. Sirois, X. Granados, T. Puig, X. Obradors,
[11] V. Matias, R.H. Hammond, YBCO superconductor wire based on IBAD-textured
A Sulfurization Method for Creating the Buffer- Layers current Flow Diverter
templates and RCE of YBCO: process economics, Phys. Procedia 36 (2012)
Architecture in REBa2Cu3O7 Coated Conductors, 36, Superconductor Science and
1440–1444.
Technology, 2023, 125005.
[12] S. Elschner, F. Breuer, M. Noe, T. Rettelbach, H. Walter, J. Bock, Manufacturing
[39] C. Lacroix, Accelerating Quench Propagation in 2G HTS Coated Conductor by
and testing of MCP 2212 bifilar coils for a 10 MVA fault current limiter, Trans.
Engineering the Tape Architecture, World Scientific Publishing, 2018,
Appl. Superconduct. 13 (2003) 1980–1983.
pp. 347–355. ISBN-10: 9813272972, chapter 16.
[13] P. Tixador, X. Obradors, R. Tournier, T. Puig, D. Bourgault, X. Granados, J.
[40] G. Deutscher, High voltage superconducting fault current limiters based on high
M. Duval, E. Mendoza, X. Chaud, E. Varesi, E. Beaugnon, D. Isfort, Quench in bulk
diffusivity substrates, J. Supercond. Novel Magn. 31 (2018) 1961.
HTS materials - application to the fault current limiter, Supercond. Sci. Technol. 13
[41] A. Saraf, B. Almog, M. Azoulay, G. Deutscher, Sapphire-Based SFCL Conductors,
(2000) 493–497.
Superconducting Fault Current Limiter, World Scientific Publishing, 2018,
[14] T. Verhaege, J.P. Tavergnier, C. Agnoux, C. Cottevieille, Y. Laumond, M. Bekhaled,
pp. 357–372 ISBN-10:9813272972, chapter 17.
P. Bonnet, M. Collet, V.D. Pham, Experimental 7.2 kVrms/1 kARMS/3 kApeak
[42] C. Lacroix, F. Sirois, Impact of current flow diverter on innovative HTS tape
current limiter system, IEEE Trans. Appl. Superconduct. 3 (1993) 574–577.
architectures for DC fault current limitation at electric fields up to 150V/m, IEEE
[15] T. Verhaege, C. Cottevieille, P. Estop, M. Quemener, J.P. Tavergnier, M. Bekhaled,
Trans. Appl. Supercond. 29 (2019), 5603407.
C. Bencharab, P. Bonnet, Y. Laumond, V.D. Pham, C. Pomarede, P.G. Therond,
[43] B. Gromoll, G. Ries, W. Schmidt, H.-P. Kraemer, B. Seebacher, B. Utz, R. Nies, H.-
Experiments with a high voltage (40kV) superconducting fault current limiter,
W. Neumueller, Resistive fault current limiters with YBCO films 100 kVA
Cryogenics 36 (1996) 521–526.
functional model, IEEE Trans. Appl. Supercond. 9 (1999) 656–659.
[16] Time Magazine, May 11, 1987, 129 (1987) No 19.
[44] C. Pop, P. Barusco, R. Vlad, K. Gupta, B. Almog, A. Saraf, G. Deutscher,
[17] N. Mitchell, J. Zheng, C. Vorpahl, V. Corato, C. Sanabria, M. Segal, B. Sorbom,
X. Granados, T. Puig, X. Obradors, High Critical Current Solution Derived
R. Slade, G. Brittles, R. Bateman, Y. Miyoshi, N. Banno, K. Saito, Anna Kario, H. Ten
YBa2Cu3O7 Films Grown on Sapphire for Current Limitation Applications, 35,
Kate, P. Bruzzone, R. Wesche, T. Schild, N. Bykovskiy, A. Dudarev, M. Mentink, F.
Superconductor Science and Technology, 2022, 054007.
J. Mangiarotti, K. Sedlak, D. Evans, D.C. Van Der Laan, J.D. Weiss, M. Liao, G. Liu,